# Authentication Source: https://docs.tryxlr8.ai/account/authentication Sign in, create your account, manage your organization, and add team members to XLR8.ai. ## Signing in Go to [app.tryxlr8.ai](https://app.tryxlr8.ai) to access XLR8.ai. On the welcome screen, choose **Sign In** or **Sign Up**. Enter your email address to sign in. You can use an email/password combination or any social login method presented on the sign-in screen. After you authenticate, XLR8.ai opens **Experiments** by default (or **Welcome** if you still need to finish setup). > [!TIP] > > Bookmark [app.tryxlr8.ai](https://app.tryxlr8.ai) for quick access. Your session persists across browser restarts unless you explicitly sign out. *** ## Creating an account **1. Open the sign-up screen** Go to [app.tryxlr8.ai](https://app.tryxlr8.ai) and click **Sign Up**. **2. Enter your details** Provide your email address and choose a password. Alternatively, use a social login option if available. **3. Verify your email** Check your inbox for a verification email. Click the link in the email to confirm your address. > [!NOTE] > > The verification email may land in your spam or promotions folder if it doesn't appear in your inbox within a few minutes. **4. Accept the terms** New accounts must accept the Terms of Service at **Onboarding → Terms** before accessing the rest of the app. **5. Set up your organization** Use the organization switcher in the header to create an organization, accept an invitation, or switch between organizations you already belong to. Organizations are how XLR8.ai groups your team and data together. *** ## Organizations XLR8.ai uses **organizations** to manage team access and data. Every experiment, content piece, and integration belongs to an organization — not to an individual user. When you sign in, you're acting on behalf of the selected organization. Switching organizations reloads the app with that organization's data. ### Creating an organization If you're the first person from your company to sign up, create a new organization from the organization switcher. You'll be assigned the **Admin** role automatically. ### Managing the organization From the organization switcher, open **Manage organization** to use Clerk's organization profile. From there you can invite members, update the organization name, and manage membership. ### Inviting team members **1. Open organization settings** Open the organization switcher in the header, then select **Manage organization**. **2. Invite members** Enter the email address of the person you want to invite and choose their role. **3. Choose their role** Select either **Admin** or **Member**. See [Organization roles](#organization-roles) below for the difference. **4. Send the invitation** Send the invitation. The invitee receives an email with a link to join your organization. Pending invitations also appear in the organization switcher. ### Organization roles XLR8.ai has two roles within an organization: > [tab] Admin > > Admins have full control over the organization. As an admin, you can: > > * Invite and remove team members > * Assign roles to other members > * Edit **Organization Brand Guidelines** > * Approve, edit, dismiss, add, import, or delete **AI Brand Learnings** > * Manage org-wide **Experiment Schedules** in Settings > * Use all product features (experiments, content, intelligence, integrations) > [tab] Member > > Members can use product features but cannot manage org-level brand guidelines, add or dismiss learnings in Settings, or change schedule settings reserved for admins. As a member, you can: > > * Create and run experiments > * Generate and optimize content > * Add comment threads as **AI Brand Learnings** from the content **Comments** sidebar > * Approve and edit **AI Brand Learnings** in Settings > * Use Insights, Action Center, Prospect Intelligence, Outreach, and Social Listening > * Configure integrations such as Webflow, WordPress, indexing, and outreach email *** ## Staying signed in and signing out XLR8.ai keeps you signed in across browser sessions by default. Your session remains active until you explicitly sign out. To sign out, click your profile avatar or name in the navigation and select **Sign Out**. You'll be redirected to the sign-in screen. *** ## Resetting your password **1. Go to the sign-in screen** Navigate to [app.tryxlr8.ai](https://app.tryxlr8.ai). **2. Click 'Forgot password'** Below the password field, click the **Forgot password?** link. **3. Enter your email** Type the email address associated with your account and submit the form. **4. Check your inbox** Open the password reset email and click the reset link. The link is time-limited — use it promptly. **5. Set a new password** Enter and confirm your new password. Once saved, sign in with the new credentials. *** ## Troubleshooting sign-in issues > [accordion] I'm not receiving the verification or password reset email > > Check your spam, junk, and promotions folders. If the email still doesn't arrive after a few minutes, make sure you entered the correct email address and try again. > [accordion] My invitation link isn't working > > Invitation links expire after a set period. Ask your organization admin to resend the invitation. Make sure you're clicking the link from the most recent invite email if you received multiple invitations. > [accordion] I'm signed in but I can't see my organization's data > > You may be signed into the correct account but not associated with the right organization. Use the organization switcher to confirm the active organization, or ask your admin to confirm your invitation was accepted. > [accordion] I forgot which email I used to sign up > > Try any email addresses you commonly use for work accounts. If you're unable to determine the correct email, contact your organization admin — they can look up members by name in the organization settings. *** ## What's next? > [card] [Quick start](/quickstart) > > Create your first experiment and start tracking your AI visibility. > [card] [Experiments](/experiments/overview) > > Learn how to run visibility and sentiment experiments across LLMs. > [card] [Settings](/integrations/settings) > > Configure your company profile, brand guidelines, and integrations. > [card] [Billing](/account/billing) > > View invoices, pay on Stripe or in XLR8, and manage your payment method. > [card] [Core Concepts](/concepts) > > Learn the key terms used across XLR8.ai. # Billing Source: https://docs.tryxlr8.ai/account/billing View invoices, pay outstanding balances on Stripe or in XLR8, and manage your payment method. **Billing** is where you review invoices, pay what's due, and manage your payment method. Navigate to **Settings → Billing** in the sidebar. *** ## Pay an invoice Invoices that are due appear under **Action required**. Each card shows card and bank transfer (ACH) amounts when both are available. You can pay in either of these ways when they appear: * **Stripe invoice** — opens Stripe's hosted invoice in a new tab. This control is shown when a Stripe invoice link is available. * **Pay on XLR8** — continues to in-app checkout. Choose **Bank transfer (ACH)** or **Credit or debit card**, then complete payment on Stripe. > [!NOTE] > > **Stripe invoice** is optional. If it is not shown, use **Pay on XLR8**. Scheduled invoices that have not been issued yet appear under **Upcoming**. *** ## Invoice history **Invoice history** lists every invoice with description, billing interval, amount, method, status, and date. * Open invoices can include a **Stripe invoice** link when a Stripe hosted invoice is available. * Paid invoices can include a **Receipt** link. If you have no invoices yet, the page asks you to contact your account manager. If the rest of the platform is locked for an outstanding invoice, you will see **Payment required**. Use **Pay now** to go to checkout, or **View billing details** to open this page. After a cancelled subscription, **Go to billing** brings you here so you can renew. *** ## Manage payment method and subscription When a payment method is on file, **Manage payment method** opens the Stripe customer portal so you can update cards or bank details. If you have an active subscription, **Cancel subscription** appears. Click it, then confirm with **Confirm cancel?** to cancel. *** ## Related > [card] [Settings](/integrations/settings) > > Configure your company profile, brand guidelines, and integrations. > [card] [Authentication](/account/authentication) > > Manage organizations, roles, and team access. # Analytics Source: https://docs.tryxlr8.ai/analytics/overview Compare experiment executions over time with trends, baselines, model compare, sources, and insights. Analytics is the multi-execution reporting view for an experiment. Use it when you want to see how visibility or sentiment changed across runs — not just the latest snapshot. Navigate to **Reporting → Analytics** in the sidebar, then select an experiment. For Google Search Console, Google Analytics, and first-touch survey data, use [Reporting](/reporting/overview) instead. *** ## Analytics vs experiment results | | Analytics | Experiment results | | - | - | - | | Scope | All successful executions over time | One selected execution | | Best for | Trends, baselines, movers | Deep dive into a single run | | Location | **Reporting → Analytics** | **Strategy → Experiments → \[experiment]** | *** ## Dashboard tabs After you open an experiment, the analytics dashboard includes: * **Overview** — KPIs, top movers, source concentration, category performance * **Trends** — charts across executions * **Model compare** — visibility experiments only * **Sources** — citation mix, domains, and tracked URLs. Sentiment **Tracked URLs sentiment** can compare to **Previous execution**, **First execution**, **Avg previous**, or a specific earlier run via **Compare with…** * **Themes** — sentiment experiments only; clustered pros and cons on a 3D/2D map, with aspect, modifier, and intensity filters * **Insights** — auto-generated insight cards *** ## Filters and baselines Use the filter bar to narrow by: * **Models** * **Categories** You can also toggle execution sanitization to include only executions where every active model returned at least one query result. On **Overview**, choose a baseline comparison: * **Default** — compare to the previous execution * **Previous** — explicit previous-execution baseline * **First** — compare against the first successful execution On sentiment experiments, category **Delta** (and **Top movers**, which use the same logic) subtracts the baseline run’s sentiment index from the current run’s index. Each index uses that run’s own queries — not only queries that appeared in both runs. If a category’s query set changed versus the baseline, the delta shows a \* and may reflect a different query mix, not only a shift in tone. *** ## Themes On sentiment experiments, the **Themes** tab groups similar pros and cons into clusters so large runs stay readable. If the map is empty, click **Sync themes**. That embeds missing phrases and clusters similar points — needed once for older runs. Use **Theme filters** to narrow phrases by **Aspect**, **Modifier**, and **Intensity**. Filters apply to the current **Pros** or **Cons** polarity — switch Pros/Cons to filter the other side. A phrase must match every dimension you set (for example, both an aspect and an intensity). Selecting every option in a dropdown is the same as leaving that dimension unfiltered. **Clear filters** removes all constraints. Filters update the cluster map, top theme lists, and introduced themes. When **Cons** is selected, they also apply to the **Cons per theme** chart and **Watched cons over time**. Theme counts show matching phrases, so mild and moderate versions of the same theme appear as separate map points. If nothing matches, the map says **No clusters match these filters**. The **Theme cluster map** opens in **3D** by default. Drag to rotate; the same color means the same cluster. Hover a point to read the full topic and phrase. Switch to **2D** to pan (drag) and zoom (Shift+drag, or the zoom controls). Use **Table** to list the same phrases by cluster and topic. * **Pros / Cons** — choose which polarity the map and filters apply to * **Outliers** — off by default; turn on to include phrases that did not join a cluster * **Cons per theme** — when Cons is selected, a bar chart shows how many unique con phrases sit in each cluster theme *** ## KPI bands **Visibility experiments** surface metrics such as overall visibility, visibility trend, winning categories, and at-risk categories. **Sentiment experiments** surface an **Overall rating** (the plurality of Positive / Neutral / Negative labels), **Negative Categories**, and brand citation share. Each response is labeled **Positive** if the score is 3 or higher, **Neutral** if it is 0 up to but not including 3, and **Negative** if it is below 0. *** ## Empty states * **No Execution Data** — run at least one successful execution before analytics can load * **Unable to Load Analytics** — retry or confirm the experiment still exists *** ## Related > [card] [Reading results](/experiments/results) > > Interpret a single execution's visibility or sentiment results. > [card] [Insights Agent](/intelligence/insights-agent) > > Ask natural language questions about experiment data. > [card] [Scheduling](/experiments/scheduling) > > Keep analytics fresh with recurring executions. > [card] [Experiments](/experiments/overview) > > Create and manage visibility and sentiment experiments. > [card] [Reporting](/reporting/overview) > > Review Search Console, Google Analytics, and first-touch survey data. # System Reporting API Source: https://docs.tryxlr8.ai/api-reference/overview Bearer-authenticated experiment and analytics endpoints for reporting integrations. The **System Reporting API** exposes read-only experiment data and analytics under `/system-api/v1/reporting/*`. It is separate from the Clerk-authenticated Platform API used by the XLR8.ai web app. Use this API when you need programmatic access to experiments, executions, raw results, analyses, citation metadata, and execution trends—for example, internal or partner reporting dashboards. ## Base URLs | Environment | Base URL | | - | - | | Production (Railway) | `https://engine-mono-production.up.railway.app` | | Production (custom domain) | `https://api.tryxlr8.ai` | Paths in this API are **absolute from the host root** (for example, `GET /system-api/v1/reporting/experiments`). ## Interactive reference * **ReDoc (production):** [engine-mono-production.up.railway.app/system-api/redoc](https://engine-mono-production.up.railway.app/system-api/redoc) * **OpenAPI JSON:** [engine-mono-production.up.railway.app/system-api/openapi.json](https://engine-mono-production.up.railway.app/system-api/openapi.json) The **API Reference** tab in this docs site is generated from the same OpenAPI document (checked into `api-reference/openapi.json`). ## Authentication Every request must include a **System Reporting API key** that XLR8 issued for your integration: ```http theme={null} Authorization: Bearer ``` You do not need a separate API-key vendor account. Store the secret like any other production credential. > [!WARNING] > > Keys are secrets. Do not embed them in client-side code or commit them to git. Reporting routes enforce an **organization allowlist** on the server. Your key must be permitted for the integration, and each experiment must belong to an allowed organization or the API returns **403 Forbidden**. Contact your XLR8 administrator for key issuance and org access. ## Endpoints (summary) | Method | Path | Purpose | | - | - | - | | `GET` | `/system-api/v1/reporting/experiments` | List experiments (filter by type, archived, sort) | | `GET` | `/system-api/v1/reporting/experiments/{experiment_id}` | Experiment detail | | `GET` | `/system-api/v1/reporting/experiments/{experiment_id}/executions` | Execution history | | `GET` | `/system-api/v1/reporting/experiments/{experiment_id}/executions/{execution_id}/results` | Raw visibility/sentiment rows | | `POST` | `/system-api/v1/reporting/experiments/{experiment_id}/executions/{execution_id}/analyse-results` | Aggregated visibility/sentiment analysis | | `POST` | `/system-api/v1/reporting/experiments/{experiment_id}/executions/{execution_id}/citation-metadata` | Citation metadata for URLs in results | | `GET` | `/system-api/v1/reporting/experiments/{experiment_id}/theme-embedding-sync-status` | Theme embedding sync status | | `GET` | `/system-api/v1/reporting/experiments/{experiment_id}/theme-cluster-tsne` | Theme cluster visualization data | | `GET` | `/system-api/v1/reporting/experiments/{experiment_id}/executions-trend` | Execution trend series (optional model breakdown) | Optional query parameters such as `models`, `breakdown_by_model`, and `sanitize_executions` are documented on each operation in the reference below. ## Example: list experiments ```bash theme={null} curl -sS \ -H "Authorization: Bearer $XLR8_REPORTING_API_KEY" \ "https://api.tryxlr8.ai/system-api/v1/reporting/experiments?archived=false&sort_by=created_at&sort_order=desc" ``` ## What is not included These routes do **not** cover: * Clerk session auth used by the web app (`/api/v1/...`) * System content generation (`POST /system-api/v1/generate-content-sections` is intentionally excluded from this OpenAPI document) For the full Platform API (Swagger), use `{base}/docs` on a running Platform API deployment. ## Errors | Status | Meaning | | - | - | | `401` | Missing or invalid API key | | `403` | Experiment outside allowed organizations | | `404` | Unknown experiment or execution | | `422` | Invalid query or path parameters | | `503` | Reporting allowlist not configured on the server | # Get Reporting Execution Results Source: https://docs.tryxlr8.ai/api-reference/system/get-reporting-execution-results /api-reference/openapi.json get /system-api/v1/reporting/experiments/{experiment_id}/executions/{execution_id}/results # Get Reporting Executions Trend Source: https://docs.tryxlr8.ai/api-reference/system/get-reporting-executions-trend /api-reference/openapi.json get /system-api/v1/reporting/experiments/{experiment_id}/executions-trend # Get Reporting Experiment Source: https://docs.tryxlr8.ai/api-reference/system/get-reporting-experiment /api-reference/openapi.json get /system-api/v1/reporting/experiments/{experiment_id} # List Reporting Executions Source: https://docs.tryxlr8.ai/api-reference/system/list-reporting-executions /api-reference/openapi.json get /system-api/v1/reporting/experiments/{experiment_id}/executions # List Reporting Experiments Source: https://docs.tryxlr8.ai/api-reference/system/list-reporting-experiments /api-reference/openapi.json get /system-api/v1/reporting/experiments # Reporting Analyse Results Source: https://docs.tryxlr8.ai/api-reference/system/reporting-analyse-results /api-reference/openapi.json post /system-api/v1/reporting/experiments/{experiment_id}/executions/{execution_id}/analyse-results # Reporting Citation Metadata Source: https://docs.tryxlr8.ai/api-reference/system/reporting-citation-metadata /api-reference/openapi.json post /system-api/v1/reporting/experiments/{experiment_id}/executions/{execution_id}/citation-metadata # Reporting Theme Cluster Tsne Source: https://docs.tryxlr8.ai/api-reference/system/reporting-theme-cluster-tsne /api-reference/openapi.json get /system-api/v1/reporting/experiments/{experiment_id}/theme-cluster-tsne # Reporting Theme Embedding Sync Status Source: https://docs.tryxlr8.ai/api-reference/system/reporting-theme-embedding-sync-status /api-reference/openapi.json get /system-api/v1/reporting/experiments/{experiment_id}/theme-embedding-sync-status # Changelog Source: https://docs.tryxlr8.ai/changelog Product updates and announcements **September 29, 2026** On generated content, expand the **Compliance report** to see **What we verified** — the checks that ran against your draft, even when it passed. The **Generation trace** now explains how the draft and source references are built, and shows a **Pipeline overview** with each phase's status. Activity-log steps use plain-language titles. When you review **Suggested changes**, each changed section opens in edit mode so you can tweak the proposed text, then **Accept** or **Reject**. To optimize a single section, click the target icon, enter a **target query**, and **Run optimize**. Apply the result or discard it — you no longer need to set a page-wide query first. Query Generation no longer accepts a sales transcript file. When your organization has analyzed calls in Prospect Intelligence, generated and imported queries can show **Sales call** source badges with the theme, type, and transcript. **September 28, 2026** **Prospect Intelligence** is now in the Strategy sidebar. Upload prospect call transcripts to see what they talked about, asked, praised, and objected to — then filter by theme or type and open quotes from the charts. In **Settings**, add **Internal Team** names so Prospect Intelligence can keep prospect turns and skip your team's speech. Company name and domain are used the same way to tell your brand from the prospect. Upload a link-placement brief PDF under **Settings → Brand Guidelines** to create a **Link plan**. On generated content, click **Apply link plan**, choose internal links and/or backlinks, then review the placements on **Suggested changes** before you accept them. On **Content Generation**, each library row now shows a shortened **content ID**. Click the copy button to copy the full ID without opening the piece. **September 25, 2026** After you read an outreach inbox conversation, you can **Mark as unread** from the inbox list or the open thread so you can come back to it later. **September 24, 2026** You can now add, edit, and remove queries on an existing experiment. Open **Settings → Queries** to review the list, then click **Edit queries** to make changes in Query Generation. Those edits apply to the next run — earlier results stay as they were. On Analytics, sentiment category **Delta** now compares each run’s own scores instead of only queries shared with the baseline. A \* next to a delta means that category’s query set changed. After you re-run a model on a completed experiment run, **Settings → Execution** now shows that model as **In progress** or **Auto-retry** instead of still showing **Succeeded**. Content generation now shows live **Generation progress** — pipeline steps, an activity log, self-review notes, and sources — instead of a generic spinner. Agentic runs can take up to about 15 minutes; keep the tab open until the draft is ready. On the generation form, turn on **Add internal links to content** to insert links to relevant pages on your company domain. When you include brand guidelines, each selected source now shows a short preview of the uploaded text. After generation, expand the **Compliance report** and **Generation trace** to review issues and how the draft was built. Sections can show **Brand learning** or **Brand guideline** badges so you can see which sources shaped them. Review brand-correction updates on the **Suggested changes** tab. Use **Regenerate with feedback** for the whole piece, or the sparkle button on a section. Accept or reject each change; unfinished reviews stay available if you leave and come back. In the Comments sidebar, click **Remember** to save selected threads as brand learnings. **Resolve All** now marks open threads as acted on without adding them to brand learnings. **September 23, 2026** Outreach **Inbox** and **Outbox** now have their own pages. Click a reply to open that conversation so you can reload it later. Mark each inbox thread **Active**, **Inactive**, or **Done**, and filter the list by those statuses. When you draft an outreach email, **Fill template** inserts the XLR8 listicle template with your sender name, company, and article already filled in. Save a **Client tagline** on **Settings → Outreach Email** so the template uses it; otherwise it keeps `[Client tagline]` for you to replace. **September 22, 2026** Experiment results now open the current run first. Tracked URL comparisons to **Previous execution** and **First execution** load afterward, so Overview and the other tabs no longer wait on every past run. On experiment **Tracked URLs impact** and **Tracked URLs sentiment**, **Avg previous** is no longer a baseline option. Use **Previous execution** or **First execution** there. Analytics **Sources** still includes **Avg previous** and **Compare with…**. **September 21, 2026** On **Settings → Outreach Email**, you can set a **Sender name** for each mailbox. Generated drafts and the From line use that name; leave it blank to derive it from the email address. When you **Generate from article**, choose **Generate as** so the draft uses that mailbox's name. After outreach is scheduled or sent, you can still edit **CC recipients** on the thread. People you add are copied on the next follow-up or reply. When you reply, **Include last message** quotes the previous email so a newly copied teammate has context. **September 17, 2026** In Query Generation, click a query's source badge to see why it was generated. When the source is an uploaded file, **Source document** links to download that document. Importing a categories-and-queries CSV now tags each query automatically. The sample and export include an optional **Citations** column; if you leave it out, tags come from your company briefing, documents, and Keyword Planner. Fill in company details first — or re-import after you do — so those tags can be enriched. **September 16, 2026** Claude Opus 4.1 is no longer available when you generate or optimize content. Use Claude Opus 4.7 or another current model instead. Content library rows, filters, and item labels now show the exact model version (for example, Claude Opus 4.7) instead of a generic family name. From a generated content piece, you can select comment threads and click **Add to brand learnings** so the feedback can be reviewed and applied to future content. Threads already saved show **Pending**, **In learnings**, or **Dismissed**. Comment thread status badges now wrap under the title so they stay fully visible in the sidebar. **September 15, 2026** When an invoice needs payment, [Settings → Billing](/account/billing) now offers **Stripe invoice** (opens Stripe's hosted invoice) alongside **Pay on XLR8** for in-app checkout. Open invoices in Invoice history can show the same Stripe link next to receipts. If Google Search Console or Bing Webmaster expires your connection, [Indexing](/indexing/overview) now shows **Reconnect required** and a **Reconnect** button so you can authorize again and restore access. After you analyze a domain in [Query Generation](/query-generation), suggested **company aliases** and **tracked competitors** appear on the Categories step for every subtype. They are saved with the query set and copied onto new experiments for mention tracking. For Competitor Alternatives and Competitor Comparison, tracked competitors also match the competitor names on that step. **September 14, 2026** New experiments now turn **Enforce Web Search** on by default when you choose GPT Instant or GPT Thinking, so those models search the web before answering unless you switch it off. **September 14, 2026** In Query Generation, queries you add or edit by hand now stay on the list when you regenerate suggestions, instead of disappearing. **September 11, 2026** On sentiment **Detailed Analysis**, the **Pros** and **Cons** filter now shows your company name instead of a generic Self label, so you can tell at a glance which phrases are about your brand versus competitors. **September 11, 2026** Keyword Planner in Query Generation now lets you choose a **Language** and one or more **Locations** for search-volume and competition data. Targeting defaults to English and the United States, and is saved with the query set so it is reused when you reopen or copy a briefing. **September 11, 2026** Sentiment **Detailed Analysis** now lets you filter **Pros** and **Cons** with a **Self** / **Competitor** toggle. Self (the default) shows phrases about your brand; Competitor shows phrases attributed to other brands, with brand labels. **Export to CSV** follows the same filter. **September 10, 2026** You can now choose how many queries to generate for each category when you create or regenerate a query set. Set **Queries per category** from 1 to 20 (default 5) on the Categories and Queries steps in Query Generation. **September 10, 2026** In **Settings → Tracked URLs**, you can paste a list of URLs (one per line or separated by commas) and add them to multiple visibility or sentiment experiments in one submit. An optional bucket applies to every URL, or XLR8.ai assigns one per URL when you leave it empty. **September 10, 2026** On Analytics **Themes**, you can now filter clustered phrases by aspect, modifier, and intensity for the current Pros or Cons view. Filters apply together at the phrase level, so theme lists and the cluster map only show matching phrases. **September 10, 2026** **Reporting** now combines Google Search Console, Google Analytics, and monthly first-touch survey data in one place. Connect Google, choose your properties, and upload a CSV per month to see search performance, site traffic, and how people say they found you. **September 10, 2026** Publishing to Webflow or WordPress now saves as a draft by default, so you can review the CMS item before it goes live. Turn on **Publish immediately** when you are ready, or use **Publish Draft** later — that action defaults to going live. Updating an existing Webflow item also keeps its current draft or live status instead of publishing it automatically. **September 9, 2026** Organization members can now approve pending **AI Brand Learnings** and edit their text in Settings. Dismissing a learning, adding one manually, importing from a document, and deleting remain limited to admins. **September 9, 2026** Query set rows on Query Generation are fully clickable — click anywhere on a row to open the set. Download CSV and archive appear when you hover a row, and those actions stay on the list. **September 8, 2026** You can now generate and optimize **PDF** content — reports, whitepapers, and other document-style pieces — then download the finished work as a letter-sized PDF from the content editor, alongside Markdown and HTML. **September 7, 2026** The theme cluster map on Analytics now opens in 3D so you can rotate clusters on larger sentiment experiments. Switch to 2D when you want to pan and zoom. When you view **Cons**, a **Cons per theme** chart shows how many unique criticisms fall in each cluster. Outliers stay hidden until you turn on **Outliers**. # Core Concepts Source: https://docs.tryxlr8.ai/concepts Key terms and concepts you need to understand how XLR8.ai works. ## Experiments An **experiment** is the core unit of work in XLR8.ai. It defines: * Your company and domain being analyzed * The type of analysis (Visibility or Sentiment) * The query categories and specific queries to test * The AI models to test against * Optional geo-targeting (country, region, city) Experiments are reusable — you run the same experiment multiple times over time to track how your AI visibility changes. ### Experiment Types > [tab] Visibility > > **Visibility experiments** measure how often your brand appears in AI responses. > > XLR8.ai submits queries to each selected AI model and checks whether your brand (and its aliases) are mentioned in the response. Results show: > > * **Mention rate** — percentage of queries where your brand appeared > * **Citation presence** — whether your website was cited as a source > * **Competitor comparison** — which other brands appeared for the same queries > > **Subtypes:** > > * `discovery` — broad discovery queries ("best project management tools") > * `informational` — informational queries about your category > * `competitor_alternatives` — queries asking for alternatives to competitors > [tab] Sentiment > > **Sentiment experiments** measure how AI models describe and perceive your brand. > > XLR8.ai collects AI responses and analyzes the tone, framing, and comparative positioning of your brand. Results show: > > * **Brand sentiment** — positive, neutral, or negative perception > * **Competitor comparison** — how you're positioned relative to competitors > * **Action items** — specific improvements to shift AI sentiment > > **Subtypes:** > > * `brand_sentiment` — how AI models describe your brand overall > * `competitor_comparison` — how AI models compare your brand to competitors ## Executions An **execution** is a single run of an experiment. Each time you click **Rerun Experiment** (or a schedule triggers), a new execution is created. Creating a new experiment also starts its first execution automatically. Executions have the following statuses: | Status | Description | | - | - | | `pending` | Execution queued, not yet started | | `in_progress` | Queries are actively being submitted to AI models | | `updating` | Results received, analysis being computed | | `success` | Execution completed successfully | | `failed` | Execution encountered an error | You can run multiple executions of the same experiment over time to build trend data and compare how your visibility changes. ## AI Models XLR8.ai supports testing against multiple AI models simultaneously. Each model represents a distinct AI search engine or assistant. Selectable models currently exposed by the frontend include: | Model ID | Display Name | | - | - | | `perplexity` | Perplexity | | `gpt-fast-gpt-5.1` | GPT Fast (GPT-5.1) | | `gpt-thinking-gpt-5.1` | GPT Thinking (GPT-5.1) | | `gpt-fast-gpt-5.3-chat-latest` | GPT Fast (GPT-5.3 Chat Latest) | | `gpt-thinking-gpt-5.4` | GPT Thinking (GPT-5.4) | | `grok-4` | Grok 4 | | `grok-4.20` | Grok 4.20 | | `gemini` | Gemini | | `claude-opus-4.8` | Claude Opus 4.8 | | `claude-opus-4.7` | Claude Opus 4.7 | | `claude-sonnet-5` | Claude Sonnet 5 | | `claude-sonnet-4.6` | Claude Sonnet 4.6 | | `claude-sonnet-4.5` | Claude Sonnet 4.5 | | `google_ai_mode` | Google AI Mode | | `microsoft_copilot` | Microsoft Copilot | > [!NOTE] > > The New Experiment picker loads models from the platform API (`/experiments/llm-models`). Display names come from the frontend model map. ## Category Map A **category map** organizes your experiment queries into labeled groups. For example: ```json theme={null} { "Project Management": [ "best project management tools for remote teams", "top task management software 2024" ], "Competitor Alternatives": [ "alternatives to Asana", "tools similar to Monday.com" ] } ``` XLR8.ai's domain analyzer suggests categories and queries automatically based on your company domain and experiment type. You build and edit this map in [Query Generation](/query-generation), then pick the saved query set when you create an experiment. On an existing experiment, organization admins can also open **Settings → Queries** and click **Edit queries** to change the linked set. ## Query sets A **query set** is a reusable briefing plus category map. It stores experiment type and subtype, company name, domain, aliases, tracked competitors, categories, queries, and optional Keyword Planner keywords with language and location targeting. Create and edit query sets under **Strategy → Query Generation**. When you create an experiment, you select a saved set instead of re-entering company details or generating queries inline. ## Company Aliases **Aliases** are alternative names, abbreviations, or spellings of your brand. XLR8.ai uses aliases when scanning AI responses to ensure it correctly identifies all mentions of your brand. For example, if your company is "Acme Corporation," you might add aliases: `["Acme", "Acme Corp", "ACME"]`. Domain analysis in [Query Generation](/query-generation) suggests aliases from your website. You can add or remove them on the Categories step. They are saved on the query set and copied onto new experiments. ## Schedules A **schedule** automates experiment executions on a recurring basis. Available frequencies include daily, weekly, bi-weekly, and monthly. Each schedule is attached to a specific experiment and triggers a new execution automatically. ## Content Sections When XLR8.ai generates content, it produces multiple **sections** — structured blocks that form the full article. Each section has: * A **name** (section title/heading) * **Content** (the written text) * A **score** — a cosine similarity score indicating how well the section's content aligns with your target AEO query Higher section scores indicate stronger alignment with the query, which tends to improve AI citation rates. ## Action Items **Action items** are specific, prioritized recommendations generated from your experiment results. They suggest concrete changes — to your content, website, or messaging — that can improve your AI visibility or shift sentiment. Action items can be marked as complete once you've addressed them. # Content generation Source: https://docs.tryxlr8.ai/content/generation Generate AEO-optimized content from scratch by targeting a specific query, content type, and LLM — then edit, reorder, and publish the result. Content generation produces structured, AEO-optimized content tailored to a query you want to rank for in AI-powered search. XLR8.ai generates the content as a set of discrete **sections**, each scored for semantic alignment with your target query. You can edit, reorder, add, or delete sections before publishing. ## Generate a new piece of content Navigate to **Content → Content Generation**, then click **Create New Content Action**. **1. Choose a content type** Select the content type that best matches the format you want to produce: | Type | When to use | | - | - | | `listicle` | "Best X for Y" or "Top N tools" queries — ranked or curated lists | | `competitor_comparison` | Side-by-side comparisons of your brand against named competitors | | `guide` | How-to or explanatory content targeting informational queries | | `ecommerce_guide` | Buying guides and product-focused content for commercial queries | | `technical_guide` | Deep-dive technical content for developer or specialist audiences | | `pdf` | Reports, whitepapers, or other document-style content intended to be shared as a PDF | > [!NOTE] > > The content type shapes the structure of the generated output. Choose the type that matches what AI models are most likely to cite for your target query. **2. Enter your target query** In the **Content Topic / Query** field, enter the exact query you want this content to rank for — for example, `best project management software for remote teams` or `how to reduce customer churn with AI`. This query is used throughout generation to focus the content and calculate section scores. Use the same phrasing your target audience would type into an AI assistant. **3. Enter company name and domain** Enter your **Company Name** and **Company Domain** (e.g., `example.com`). XLR8.ai uses these to personalize the content and position your brand within the generated output. > [!TIP] > > If you've saved your company details in **Settings**, they are pre-filled automatically. **4. Add a content title** Enter a **Content Title** — the headline that will appear at the top of the generated piece. This becomes the H1 of your content and is factored into AEO scoring. **5. Optionally add competitors and context** Two optional fields let you guide the output: **Competitors** (available for `listicle` and `competitor_comparison` types): Add competitor brand names by typing each name and pressing **Enter** or **,**. XLR8.ai incorporates these into the comparison structure. **Additional context**: Paste in brand voice, product names, target audience details, or any other preferences that should shape the content — for example: > Brand voice: Friendly and authoritative. Products: AI-powered analytics suite. Target audience: Marketing leaders at B2B SaaS companies. If you have generated content before, a **Previous additional context** control appears above the text field. Use it to reuse context from an earlier piece: * Leave it on **Add new context** to type fresh instructions. * Open the list to pick a previous context (shown with title, date, query, and a short preview). * When you have more than five saved contexts, search by title, query, or text. * Use the preview (eye) icon to read the full context, then click **Use this context**. The control is hidden when there are no reusable previous contexts. > [!NOTE] > > Some older saved contexts already include a brand guide. Those items show a **Legacy context includes its original brand guide** notice. If you reuse one, your current brand-guide selections are not added again. **6. Optionally add links** Two toggles control extra linking during generation: * **Add backlinks to content** — insert third-party authoritative source links. * **Add internal links to content** — insert links to relevant pages on your company domain from the indexed sitemap. This option is available once a company domain is set. **7. Choose an LLM model** Select the LLM you want to use for generation. The content form currently offers: * GPT-5.1 * GPT-5.4 * Claude Opus 4.7 * Claude Sonnet 4.5 * Claude Sonnet 4.6 * Gemini 3 Flash Preview * Gemini 3.1 Pro Preview * Grok 4 * Grok 4.20 Different models have different strengths. If you are generating content targeting a specific AI audience, consider generating with that family of models to better match its citation patterns. **8. Generate and review sections** Click **Generate Content**. For most content types, XLR8.ai runs an agentic workflow and shows **Generation progress** instead of a spinner — a **Pipeline overview** (write full draft including source references, optional links, compliance report, and save), a detailed activity log with readable step titles, self-review notes, and sources applied so far. Agentic runs can take up to about 15 minutes. Keep the tab open; XLR8.ai keeps polling until the draft is ready. If you open the piece from the library while it is still running, you see the same progress view. Once complete, you are taken to the content detail view where you can review, edit, and manage each section. ## Working with sections Generated content is structured as a series of **sections**, each with a name, body text, and a section score. Sections map to the logical parts of your content — introduction, key points, comparisons, conclusions, and so on. ### Understanding section scores Each section displays a score between 0 and 1 representing its cosine similarity to your target query: * **0.7–1.0** — Strong alignment. The section is directly relevant to the query and likely to be cited. * **0.4–0.69** — Moderate alignment. The section supports the piece but may need tightening. * **0.0–0.39** — Low alignment. Consider rewriting or removing the section. Scores update automatically when you edit a section's content, giving you live feedback as you refine the text. > [!TIP] > > Focus your editing on structurally important sections with low scores — the introduction, primary claims, and conclusion. Sections that score low and are non-essential are good candidates for deletion. ### Editing sections Click into any section's text area to edit it inline. Your changes are saved automatically. The score recalculates after you finish editing. ### Optimizing a section Click the target icon on a section to optimize that section for a specific query. Enter a **target query** (or keep the default from the piece), then click **Run optimize**. Review the results and click **Apply optimized** to save the new text, or **Discard** to keep the original. You do not need to set a page-wide query first. > [!NOTE] > > This optimizes for semantic similarity to the query, which can make the wording less conversational. Review the result before you apply it. ### Adding sections Use the **insert** control between any two sections to add a new section at that position. You can also append a section at the end of the document. New sections start with a score of 0 and recalculate once you add content. ### Reordering sections Drag sections into a new order using the drag handle on the left side of each section card. Reordering does not affect individual section scores. ### Deleting sections Click the delete icon on any section to remove it permanently. You cannot undo a section deletion, so review before removing. > [!WARNING] > > Deleting a section is permanent. If you want to keep the content for reference, copy the text before deleting. ## Metadata description The first section of your content includes a **metadata description** field. This is the meta description that gets published alongside your content when you push it to Webflow. Keep it concise (under 160 characters) and include your target query keywords. ## Brand guidelines If your organization has configured **Organization Brand Guidelines** in Settings, content generation can include them by default. You can also toggle **Use org brand guidelines** on the generation form. When the toggle is on, each guidelines text and uploaded file shows a short preview of the uploaded text so you can choose what to include. When you upload a **link-placement brief** PDF under **Settings → Brand Guidelines**, XLR8.ai parses it into a **Link plan**. Those plans appear on the same Settings card. They are not inserted automatically during generation — apply them later from the content toolbar. Approved **AI Brand Learnings** from comment threads further steer future content. Use **Remember** in the Comments sidebar to save selected threads for review in Settings. ## Comments and review On the content detail view, the **Comments** sidebar lists comment and suggestion threads on the generated piece. A **Suggested changes** tab sits next to Comments for brand-correction and link-plan diffs. * Filter by type (**Comments**, **Suggestions**, or **All types**) and status (**Open**, **Resolved**, or **All**). * Select a thread to reply, **Resolve** it, or **Reopen** it. * Pending suggestions also offer **Accept** and **Reject**. Accepting a suggestion applies the edit and marks the thread as acted on. * When the status filter is **Open**, click **Resolve All** to mark every open thread as acted on. The confirmation notes that this does not add brand learnings. Thread actions are disabled while bulk resolve is in progress. * Check the threads you want to reuse, then click **Remember**. The button shows how many threads are selected. Threads already saved as a brand learning cannot be selected again. * Saved threads show a status badge under the title: **Pending**, **In learnings**, or **Dismissed**. Selected threads become **AI Brand Learnings** in [Settings](/integrations/settings) for review. Threads that are already saved or not reusable are skipped. ## Reviewing a generated draft After generation finishes, the detail view can include: * **Compliance report** — expand it to see whether the draft passed, is blocked, or has issues. A **What we verified** list shows which checks ran. Issues are grouped by category and can include a short fix hint. Prompt contradictions appear separately when the draft conflicts with your instructions. * **Generation trace** — expand it to read how the draft and source references were built, a **Pipeline overview** with each phase's status, the detailed activity log, and self-review notes. Long notes have **View full note**. * **Brand learning** or **Brand guideline** badges on sections — hover or click to see which approved learnings or guidelines shaped that section. ## Suggested changes Use brand-correction feedback or a link plan to update a generated piece without rewriting it by hand. * Click **Regenerate with feedback** in the toolbar to apply approved brand learnings across the piece. * Click **Apply link plan** to insert internal and backlink placements from a plan in Settings. Choose a plan, then turn **Internal links** and **Backlinks** on or off before you apply. * Or click the refresh button on a section to regenerate only that section. * Review the result on the **Suggested changes** tab. Changed sections open in edit mode with the suggested diff so you can tweak the proposed text, then **Accept** or **Reject**. * If you leave the page before you finish reviewing, your pending suggested changes are saved and restored when you return. If **Apply link plan** shows no plans, upload a link-placement brief PDF under **Settings → Brand Guidelines**. ## Download and export From the content editor, click the download button to export the full piece as **Markdown**, **HTML**, or **PDF**. The PDF uses your current section text — including any optimized versions you applied — and saves a letter-sized file named after the content title. ## After generating: publish Once you are satisfied with your sections, publish from the content toolbar to: * **Webflow** — see [Webflow integration](/integrations/webflow) * **WordPress** — see [WordPress integration](/integrations/wordpress) The first publish defaults to a draft. Turn on **Publish immediately** if you want the item live right away, or use **Publish Draft** later from the toolbar. After the first Webflow publish, XLR8.ai tracks whether your local content has changed since the last sync. ## Viewing your content library All generated content is accessible from **Content → Content Generation**. The library shows: * Content title and a short preview from the first section * Content type and the exact LLM version used (for example, Claude Opus 4.7) * Creation date * A shortened **content ID**, with a copy button for the full ID * Webflow publish status (if connected) Click any row to open the full content editor. Copying the ID stays on the list — it does not open the piece. > [card] [Content optimization](/content/optimization) > > Rewrite existing content for better AEO alignment instead of generating from scratch. > [card] [Webflow integration](/integrations/webflow) > > Publish generated content directly to your Webflow CMS. > [card] [WordPress integration](/integrations/wordpress) > > Publish posts or pages to connected WordPress sites. > [card] [Videos](/content/videos) > > Turn a script or content piece into an explainer video. # Content optimization Source: https://docs.tryxlr8.ai/content/optimization Rewrite existing content for better AEO alignment. Paste your text or provide a URL and XLR8.ai returns an optimized version with structured edit instructions. Content optimization takes content you already have — a published blog post, a landing page, a draft — and rewrites it to better align with a specific AEO query. XLR8.ai returns an optimized version of your content alongside a set of annotated **edit instructions** that explain what was changed and why. Use optimization when you want to improve an existing page rather than start from scratch, or when experiment results reveal a high-value query that your current content doesn't adequately target. > [!TIP] > > After reviewing your experiment results, note the queries where competitors appear but you do not. Those are the highest-value targets for content optimization. ## How to optimize content Navigate to **Content → Content Optimizer**, then click **Open Content Optimizer**. **1. Choose your content source** Select how you want to provide the content to optimize: > [tab] Direct content > > Select **Direct Content** and paste your text into the content field. Use this option for drafts, copied web page text, or any content you have locally. > > XLR8.ai accepts plain text or markdown. Content is processed as-is, so strip out navigation, footers, and other non-body text before pasting. > [tab] Content URL > > Select **Content URL** and enter a fully qualified URL (starting with `https://`). XLR8.ai fetches the page, extracts the readable body text, and uses that as the optimization input. > > Use this option when you want to optimize a live published page without manually copying its content. > > > [!NOTE] > > > > XLR8.ai extracts up to 50,000 characters of readable text from the fetched page. Very long pages are truncated at that limit. The URL must be publicly accessible — pages behind authentication are not supported. **2. Enter your target AEO query** In the **Content Topic / Query** field, enter the query you want the optimized content to rank for — for example, `enterprise data governance best practices` or `how does RAG work in LLMs`. The optimizer rewrites your content to maximize semantic alignment with this query. Use precise, natural-language phrasing that matches how your audience would query an AI assistant. **3. Enter company name and domain** Enter your **Company Name** and **Company Domain**. These are used to preserve brand references and domain-specific framing during optimization. **4. Select a content type** Choose the content type that best describes your input content: | Type | Use when your content is… | | - | - | | `listicle` | A ranked list or curated "best of" piece | | `competitor_comparison` | A side-by-side product or brand comparison | | `web_page` | A marketing or product page | | `guide` | A how-to or explanatory article | | `ecommerce_guide` | A buying guide or product-focused piece | | `technical_guide` | A technical deep-dive for specialist readers | | `pdf` | A report, whitepaper, or other document intended to be shared as a PDF | The content type informs how the optimizer structures and frames the rewritten output. **5. Choose an LLM model** Select the model to use for optimization. The content form currently offers: * GPT-5.1 * GPT-5.4 * Claude Opus 4.7 * Claude Sonnet 4.5 * Claude Sonnet 4.6 * Gemini 3 Flash Preview * Gemini 3.1 Pro Preview * Grok 4 * Grok 4.20 If no model is specified, XLR8.ai uses the platform default. **6. Optionally add context** Use the **Additional Context** field to provide constraints or preferences that should be respected during optimization — for example: > Preserve the brand voice: direct and technical. Do not remove the section on security compliance. Priority keyword: "zero-trust architecture". This is also the place to flag content you want to keep unchanged, terminology that must stay consistent, or structural preferences. **7. Run optimization and review the output** Click **Optimize Content**. Optimization typically takes 15–45 seconds. When complete, the results panel appears showing the optimized content and edit instructions side by side. ## Understanding optimization results The results view has two panels: ### Optimized content The left panel displays the full rewritten content. You can copy it to the clipboard using the copy button in the header. If you click an edit instruction in the right panel, the corresponding change is highlighted in yellow within the optimized content, and the view scrolls to that location. This makes it easy to review each change in context. ### Edit instructions The right panel lists the structured edit instructions that accompany the optimization. Each instruction includes: * **Action** — one of: `Add`, `Replace`, `Remove`, `Move`, or `Rewrite` * **Instruction** — a human-readable description of the change made and why Click any instruction card to highlight the relevant passage in the optimized content panel. Use the edit instructions to understand the reasoning behind each change, and to decide which edits to apply if you are manually updating your content rather than using the output as-is. > [!NOTE] > > Optimization results are not saved to your content library. Copy the optimized content before navigating away, or apply the edits directly to your source page. ## Content alignment analysis In addition to the full optimizer, XLR8.ai includes a **Content Alignment** tool under **Content → Additional Tools**. Use content alignment to audit how well a piece of existing content aligns with a query — without rewriting it. Content alignment returns: * A **document score** (0–1) representing overall query alignment * **Chunk scores** for each segment of the content, showing which parts align well and which don't * **Statistics** on how many chunks are high-performing vs. low-performing * An **audit** covering URL structure, title and meta description length and keyword presence, H1 structure, heading nesting, and word count ### When to use content alignment vs. content optimization | Use content alignment when… | Use content optimization when… | | - | - | | You want to audit a page without changing it | You are ready to produce an improved version | | You need a quick score to prioritize which pages to update | You want structured, actionable edit instructions | | You're comparing multiple pages against the same query | You want AI to do the rewriting work for you | > [!TIP] > > Run a content alignment check first to see the current score, then run optimization on the pages with the lowest scores and highest strategic value. This gives you a clear before/after benchmark. ## Next steps > [card] [Content generation](/content/generation) > > Generate net-new AEO-optimized content from scratch instead of optimizing existing content. > [card] [Experiments](/experiments/overview) > > Identify which queries to target by running visibility and sentiment experiments. > [card] [Publish to Webflow](/integrations/webflow) > > Publish generated content directly to your Webflow CMS after optimizing. > [card] [Action center](/intelligence/action-center) > > See prioritized recommendations that surface the highest-impact queries to target. # Content Source: https://docs.tryxlr8.ai/content/overview Generate AEO-optimized content from scratch and improve existing pages so AI search engines recommend your brand more often. The Content section of XLR8.ai helps you close the loop between insight and action. Identify high-value queries through your experiments, then produce or improve content that targets them — and publish it to your CMS. > [card] [Content generation](/content/generation) > > Generate new AEO-optimized content targeting a specific query, content type, and LLM. > [card] [Content optimization](/content/optimization) > > Rewrite an existing page or draft to better align with target queries. > [card] [Videos](/content/videos) > > Generate explainer videos from a script or an existing content piece. > [card] [Publish](/integrations/webflow) > > Publish generated content to Webflow or WordPress. ## Content types When generating or optimizing content, choose a content type that shapes the structure of the output: | Content type | Description | | - | - | | Listicle | A ranked or curated list — ideal for "best X" and "top Y" queries | | Competitor Comparison | A side-by-side analysis of your brand versus named competitors | | Guide | A how-to or explanatory guide targeting an informational query | | Ecommerce Guide | A buying or product guide optimized for commercial queries | | Technical Guide | A deep-dive technical document targeting specialist queries | | PDF | A report, whitepaper, or other downloadable document | Choose the type that matches the format AI models are most likely to cite for your target query. For example, if your query is "best project management tools for startups", a listicle or competitor comparison is usually a better fit than a general guide. ## Brand guidelines and learnings Content tools can use: * **Organization Brand Guidelines** from Settings (text and PDFs) * **Link plans** parsed from link-placement PDFs — apply them from the generated content toolbar * **AI Brand Learnings** approved from comment threads you add in the content editor Configure these under [Settings](/integrations/settings). ## How content scoring works Generated sections can receive a **section score** — a cosine similarity value that measures how closely the section's text aligns with your target query. * **Higher scores** indicate stronger semantic alignment with the query * **Lower scores** flag sections that drift from the query focus Use scores to prioritize editing effort on structurally important sections with weak alignment. > [!NOTE] > > Section scores measure query alignment, not overall writing quality. A section can be well-written but score low if it covers tangential topics. ## Managing your content library All content you generate is saved to your organization's content library at **Execution → Content Generation**. From the library you can: * Open any piece of content and edit sections * Copy a content ID from a row without opening the piece * Review comments and resolve threads * Apply a link plan and review suggested placements * Publish to Webflow or WordPress from the content toolbar * Delete content you no longer need ## Optional Additional Tools If you enable **Show Additional Tools in the sidebar** in Settings, you also get: * **Content Alignment** — score content chunks against a query or URL * **Cosine Similarity Maximiser** — iterative similarity optimization * **AEO Audit** — homepage AI discoverability scorecard See [AEO Audit](/intelligence/aeo-audit) for details. ## Next steps > [card] [Generate content](/content/generation) > > Step-by-step guide to creating a new AEO-optimized content piece. > [card] [Optimize content](/content/optimization) > > Learn how to rewrite an existing page for better query alignment. > [card] [Videos](/content/videos) > > Turn scripts into explainer videos. > [card] [Run experiments](/experiments/overview) > > Identify which queries to target before generating content. # Videos Source: https://docs.tryxlr8.ai/content/videos Generate YouTube explainer videos from a custom script or an existing content piece. Use **Videos** to generate explainer videos from a script you write or from an existing content piece in XLR8.ai. Navigate to **Execution → Videos** in the sidebar. *** ## Create a video **1. Open New video** From the Videos list, click **New video**. **2. Provide a script** Optionally **Pull from existing content piece**, or write a custom script. Required fields: * **Video title** * **Script** — for a roughly 5-minute video, aim for about 700–800 words **3. Choose a voice** Select a **Voice** from the available VideoGen voices. **4. Configure video settings** Set: * **Aspect ratio** — 16:9 Landscape, 9:16 Portrait, or 1:1 Square * **Visual pacing** — Slow, Medium, or Fast * **Quality** — Low, Standard, or High * **Language** * **Visual style** — Stock, AI Image, or Entity * **Voice speed** — 0.5× to 2.0× * **Production notes for AI** (optional) **5. Generate** Click **Generate video**. The job is submitted to VideoGen and appears in your Videos list with status updates. *** ## Job statuses | Status | Meaning | | - | - | | Pending | Queued for processing | | Processing | Video is being generated | | Completed | Ready to preview and download | | Failed | Generation failed — open the job for details | On a completed job, open the detail page to preview the video or **Download MP4**. > [!NOTE] > > The Videos page currently offers **Explainer video** generation. Avatar video options are not enabled in the product UI. *** ## Related > [card] [Content generation](/content/generation) > > Create the content piece you can pull into a video script. > [card] [Content overview](/content/overview) > > See all content tools available in XLR8.ai. # Experiments Source: https://docs.tryxlr8.ai/experiments/overview Track your brand's AI presence and sentiment by running experiments across multiple LLMs. Learn how to create, manage, and act on experiments in XLR8.ai. An **experiment** is the core unit of measurement in XLR8.ai. Each experiment defines what you're tracking — your brand, the type of analysis, the queries to test, and the AI models to run against. You run experiments repeatedly over time to see how your AI visibility and brand sentiment change. > [card] [Query Generation](/query-generation) > > Build reusable query sets — company briefing, categories, and queries — then pick a set when you create an experiment. > [card] [Visibility experiments](/experiments/visibility) > > Measure how often your brand appears in AI-generated responses across leading models. > [card] [Sentiment experiments](/experiments/sentiment) > > Understand how AI models describe and perceive your brand in their responses. > [card] [Reading results](/experiments/results) > > Interpret mention rates, sentiment scores, citations, and AI-generated analysis. > [card] [Analytics](/analytics/overview) > > Compare executions over time with trends, baselines, and source concentration. > [card] [Scheduling](/experiments/scheduling) > > Automate recurring experiment runs so your data stays current without manual effort. ## Visibility vs. sentiment — which should you use? > [tab] Visibility > > Choose a **Visibility Analysis** experiment when you want to measure whether your brand is being mentioned in AI responses at all. > > Visibility answers questions like: > > * "Does Perplexity mention us when users search for our category?" > * "How often do we appear vs. our competitors?" > * "Is our website being cited as a source?" > > Visibility results show mention rates per model, citation presence, and competitor comparison — all broken down by query category. > > **Best for:** Brands establishing an AI presence baseline, competitive benchmarking, and tracking the impact of content changes over time. > [tab] Sentiment > > Choose a **Sentiment Analysis** experiment when you want to understand the tone and framing of how AI models talk about your brand. > > Sentiment answers questions like: > > * "Does ChatGPT describe us positively or negatively?" > * "How do AI models position us relative to a competitor?" > * "What pros and cons do AI models associate with our brand?" > > Sentiment results include sentiment scores, listed pros and cons per query, and competitor positioning data. > > **Best for:** Brand reputation monitoring, competitive positioning analysis, and identifying specific messaging improvements to address. ## Creating a new experiment Company details, experiment type, and queries now live on a **query set**. Create or edit those in [Query Generation](/query-generation) first. Creating an experiment then has two steps: **Query set** and **Create**. **1. Create a query set (if you do not have one)** From the sidebar under **Strategy**, click **Query Generation**, then **New query set**. Enter company details, analyze the domain, review categories, aliases, and tracked competitors, generate or import queries, and **Save query set**. See [Query Generation](/query-generation) for the full walkthrough, including CSV import, Keyword Planner, and comments. **2. Open New Experiment** From the sidebar under **Strategy**, click **Experiments**. Then click **New Experiment**. **3. Step 1 — Query set** Pick a saved query set that already has queries. Experiment type, company, domain, aliases, and tracked competitors come from that set — you will not enter them again. The selector shows each set's name, company, category count, and query count. After you select one, a summary shows company, domain, type/subtype, aliases, tracked competitors, and query counts. Use **Edit** to open the set in Query Generation, or **Manage query sets** to go to the list. If you have no saved sets with queries yet, create one in Query Generation first. Click **Continue**. **4. Step 2 — Create** Confirm the **Experiment Summary** — including company aliases and tracked competitors when the query set has them. Those names are saved on the experiment for competitor mention tracking. Then configure: * **Model Selection** — choose which AI models to test. If you select GPT Instant or GPT Thinking, an **Enforce Web Search** switch appears. New experiments start with this on, so those models search the web before answering unless you turn it off. * **Location Information** (optional) — country, region/state, and city for geo-targeted answers Click **Create Experiment**. XLR8.ai saves the configuration, opens the experiment detail page, and starts the first execution automatically. ## Managing your experiments ### Filtering and sorting On the **Experiments** page, search experiments and filter by type, subtype, or archival status. Sort by creation date or name. ### Pinning Pin important experiments to keep them at the top of your list. Click the pin icon on any experiment card to toggle the pinned state. ### Tags Add tags to categorize and filter your work. Tags are editable from the experiment detail view. ### Archiving Archive experiments you no longer actively track. Archived experiments are hidden from the default view but remain accessible via the archived filter. Archiving does not delete any data. ### Rerunning and scheduling From an experiment detail page you can: * **Rerun Experiment** — start a new execution with the current configuration * **Schedule** — set a recurring cadence (see [Scheduling](/experiments/scheduling)) To re-run a single model on the current execution instead of starting a new run, use **Settings → Execution**. ### Experiment settings Open **Settings** on an experiment to manage: * **Execution** — per-model status for the selected run (**Succeeded**, **Failed**, **In progress**, or **Auto-retry**). Organization admins can click **Re-run** on a model, or **Re-run failed** when more than one failed. That writes into the same execution, and the row shows **In progress** or **Auto-retry** while it runs — even if the overall execution already completed. * **Web search** — organization admins can toggle **Enforce web search**. When on, supported GPT models search the web before responding. New experiments start with this on. The change saves immediately and applies to future runs, not past executions. * Company aliases — copied from the query set; you can still edit them here * Competitors — pre-filled from the query set's tracked competitors; edit names and aliases here * Tracked URLs — or add the same URLs to several experiments from **Settings → Tracked URLs** * **Queries** — the **Queries** tab lists every query by category. Organization admins can click **Edit queries** to open the linked set in [Query Generation](/query-generation), where you can add, edit, or remove queries. Changes apply to the next run; earlier execution results stay as they were. If the experiment is not linked to a query set yet, **Edit queries** creates one from the current queries. * Categories * Outreach excluded domains (visibility experiments) ### Deleting Delete an experiment permanently from the experiment detail view. This removes the experiment and all of its executions and results. > [!WARNING] > > Deleting an experiment is irreversible. All execution history and results are permanently removed. Archive the experiment instead if you may need the data later. ## What's next? > [card] [Query Generation](/query-generation) > > Build and edit the query sets you pick when creating an experiment. > [card] [Visibility experiments](/experiments/visibility) > > Configure tracked URLs, company aliases, geo-targeting, and model selection for visibility. > [card] [Sentiment experiments](/experiments/sentiment) > > Understand sentiment subtypes, action item sync, and how to interpret sentiment scores. > [card] [Analytics](/analytics/overview) > > Track how visibility and sentiment change across executions. > [card] [Scheduling](/experiments/scheduling) > > Set up automatic recurring executions to track changes over time. # Reading results Source: https://docs.tryxlr8.ai/experiments/results Understand how to read execution results, compare models, use AI-generated analysis, track trends, add notes, and explore citation metadata. After an execution reaches `success` status, XLR8.ai displays a full results view for that execution. This page explains how to read and work with results for both visibility and sentiment experiments. ## Navigating executions An experiment can have many executions — one for each time you (or a schedule) triggered a run. From the experiment detail page: * The **execution list** shows all executions, sorted most-recent first, along with their status and creation date. * Click any execution to make it the active view. The results panel updates to show that execution's data. * Only executions with `success` status have results available. Executions in `pending`, `in_progress`, `updating`, or `failed` states display their current status instead. The most recent successful execution is selected by default when you open an experiment. ## Filtering by model If your experiment runs across multiple AI models, you can filter the results view to show data for one or more specific models. Use the **model filter** in the results toolbar to select the models you want to compare. This is useful when you want to isolate how a single model treats your brand, or when comparing two models side by side without the noise of all others. ## Visibility results Visibility experiment detail pages use three tabs: * **Overview** — high-level mention and competitive summary for the selected execution * **Query Analytics** — query-level breakdowns across models and categories * **Source Analytics** — citation and source-domain analysis, including tracked URL impact when configured Use the execution selector and model filter in the toolbar to change which run and models you are viewing. If you've configured [tracked URLs](/experiments/visibility#tracked-urls), Source Analytics includes **Tracked URLs impact** — citation counts and whether your brand was mentioned on those citing prompts. When earlier successful runs exist, choose **Previous execution** or **First execution** as the baseline. Those comparisons load after the rest of the page; the heading stays visible while they do. The first successful run still shows the table, without baseline buttons. **Avg previous** is not available here — use [Analytics → Sources](/analytics/overview) when you want an average of prior runs. ## Sentiment results Sentiment experiment detail pages use two tabs: * **Overview** — overall sentiment summary for the selected execution. The overall score uses a −5 to +5 scale, labeled **Negative** (below 0), **Neutral** (0 to less than 3), or **Positive** (3 to 5). **Positive Features** counts categories whose mean score is Positive (≥ 3). **Most Positive** and **Most Negative** are the highest- and lowest-scoring categories in this run. * **Detailed Analysis** — category-level sentiment, pros/cons, and supporting citations. Filter rows by polarity: **Positive (3.0 to 5.0)**, **Neutral (0.0 to less than 3.0)**, or **Negative (below 0.0)**. On the **Pros** and **Cons** columns, use the company-name / **Competitor** toggle (defaults to your company name) to show phrases about your brand or phrases attributed to other brands. Competitor view labels each phrase with the brand. **Export to CSV** uses the same filter and records the selected mode. Phrases without a brand attribution are hidden in both views. If the experiment has tracked URLs, **Detailed Analysis** shows **Tracked URLs sentiment** — mean sentiment on queries that cite those URLs. Compare change to **Previous execution** or **First execution** the same way as visibility Source Analytics. **Avg previous** is on [Analytics → Sources](/analytics/overview), not on the experiment page. Review repeated cons across categories and models — those are usually the highest-priority perception issues to address. Use **Sync Action Items** when available to send them to the [Action Center](/intelligence/action-center). ## AI-generated analysis On any successful execution, click **Analyse Results** to generate an AI-powered summary of the execution's findings. For **visibility experiments**, the analysis includes: * A narrative summary of your brand's visibility across models and categories * Identification of your strongest and weakest performing query categories * Competitor brand presence patterns * Tracked URL impact (if tracked URLs are configured) For **sentiment experiments**, the analysis includes: * An overall assessment of brand sentiment across the execution * The most frequently cited pros and cons * Patterns in how different models frame your brand * Key areas for improvement > [!NOTE] > > AI analysis is generated on demand. It is not cached — each time you click **Analyse Results**, XLR8.ai generates a fresh analysis based on the current results data and your active model filter. ## Execution notes You can attach notes to any execution to record context — such as what content changes you made before a run, external factors that might have affected results, or observations you want to preserve. ### Adding a note 1. Open the execution in the results view. 2. Click **Add Note** in the execution notes section. 3. Enter your note text and save. Notes are displayed in chronological order on the execution and help you correlate later Analytics trends with real-world changes. ### Editing a note 1. Find the note in the execution notes list. 2. Click the edit icon next to the note. 3. Update the text and save. ### Deleting a note 1. Find the note in the execution notes list. 2. Click the delete icon next to the note. 3. Confirm the deletion. > [!WARNING] > > Deleting a note is permanent. There is no way to recover a deleted note. ## Trend view The **trend view** compares results across all successful executions for an experiment, displayed as a timeline. Use it to track how your visibility or sentiment changes over time. For **visibility trends**, the chart shows mention rate per execution, with breakdowns by model available. You can spot the impact of content updates or competitor activity as inflection points in the trend line. For **sentiment trends**, the chart shows sentiment score per execution. Declining scores can signal that new sources — competitor content, negative reviews — have entered the AI training or retrieval pool. Each data point on the trend chart corresponds to a specific execution. Hover over a data point to see: * The execution serial number and date * The computed metric value for that execution * Any notes attached to that execution > [!TIP] > > Add an execution note whenever you publish new content, make significant website changes, or observe an unexpected shift in results. Notes on the trend chart make it easy to correlate actions with outcomes. ## Citation metadata For a deeper look at citations, use the **citation metadata** view. This enriches each citation URL with: * **Page title** — the title of the cited page * **Page description** — the meta description or summary of the cited page * **Query context** — which query and model produced this citation Citation metadata helps you understand: * What types of content AI models are drawing on (blog posts, review sites, product pages, competitor content) * Whether your own cited pages are well-optimized for the queries that surfaced them * Which third-party sites have influence over AI model responses in your category > [card] [Visibility experiments](/experiments/visibility) > > Configure tracked URLs, aliases, and geo-targeting for your visibility experiments. > [card] [Sentiment experiments](/experiments/sentiment) > > Understand sentiment subtypes and action item sync for sentiment experiments. > [card] [Analytics](/analytics/overview) > > Compare executions over time with trends, baselines, and insights. > [card] [Scheduling](/experiments/scheduling) > > Automate executions so your trend data stays current over time. > [card] [Action center](/intelligence/action-center) > > Turn sentiment results into prioritized action items. # Scheduling Source: https://docs.tryxlr8.ai/experiments/scheduling Automate recurring experiment executions so your AI visibility and sentiment data stays current without manual effort. A **schedule** automates the execution of an experiment on a recurring basis. Instead of manually clicking **Rerun Experiment** each time, a schedule triggers a new execution at a fixed interval — so you always have fresh data to track how your AI presence changes over time. Schedules are attached to a specific experiment. When a schedule fires, it creates a new execution with the same configuration as the experiment: same queries, same models, same geo-targeting. ## Schedule frequencies When you create a schedule, choose one of these cadences: * **Daily** * **Weekly** * **Bi-weekly** * **Monthly** ## Creating a schedule **1. Open the experiment** Navigate to **Experiments** and open the experiment you want to schedule. **2. Open the schedule settings** In the experiment detail view, click **Schedule** (or the schedule icon in the header). If no schedule exists yet, you'll see an option to create one. **3. Choose a frequency** Select how often you want the experiment to run automatically: **Daily**, **Weekly**, **Bi-weekly**, or **Monthly**. **4. Set the first run date** Pick a **next run date** — the date and time when the schedule should first trigger. All subsequent runs will be calculated from this anchor date based on the frequency you selected. **5. Save the schedule** Click **Create Schedule**. The schedule is now active and linked to the experiment. You can see the next scheduled run date in the experiment detail view. ## Updating a schedule To change the frequency or next run date of an existing schedule: 1. Open the experiment detail view. 2. Click the schedule settings. 3. Update the **frequency** or **next run date** as needed. 4. Save the changes. The updated schedule takes effect immediately. The next execution will fire at the new next run date. ## Pausing a schedule To temporarily stop a schedule without deleting it, toggle the schedule to **inactive** in the schedule settings. The schedule remains attached to the experiment but will not trigger new executions until you reactivate it. ## Deleting a schedule To permanently remove a schedule: 1. Open the experiment detail view. 2. Open the schedule settings. 3. Click **Delete Schedule** and confirm. The experiment remains intact — only the automation is removed. You can still run executions manually. > [!NOTE] > > Deleting an experiment also deletes its associated schedule. You do not need to delete the schedule separately before deleting the experiment. ## Viewing scheduled execution history All executions — whether triggered manually or by a schedule — appear in the same execution list on the experiment detail page. There is no separate view for scheduled runs. Use [Analytics](/analytics/overview) to compare results across all executions over time, regardless of how they were triggered. Organization admins can also manage schedules in bulk from **Settings → Experiment Schedules**. > [card] [Experiments overview](/experiments/overview) > > Learn how to create and manage experiments on XLR8.ai. > [card] [Analytics](/analytics/overview) > > Compare scheduled execution results over time. > [card] [Concepts](/concepts) > > Understand how schedules, executions, and experiments relate to each other. > [card] [Settings](/integrations/settings) > > Manage org-wide experiment schedules from Settings. # Sentiment experiments Source: https://docs.tryxlr8.ai/experiments/sentiment Analyze how AI models describe and perceive your brand. Learn about sentiment subtypes, scores, pros and cons, and action item sync. A **sentiment experiment** tells you *how* AI models talk about your brand — not just whether they mention you. XLR8.ai submits your queries to each selected AI model, collects the responses, and analyzes the tone, framing, pros and cons, and comparative positioning of your brand in those responses. You choose the sentiment subtype and generate or import queries in [Query Generation](/query-generation), then pick that query set when you create the experiment. Domain analysis suggests company aliases and tracked competitors, which are saved with the set and copied onto the experiment. Run sentiment experiments to answer questions like: * "Does ChatGPT describe our product positively or negatively?" * "What reasons do AI models give when recommending us — or not recommending us?" * "How do AI models position us relative to a specific competitor?" * "What are the most common criticisms AI models associate with our brand?" ## Subtypes > [accordion] Brand sentiment > > **Subtype:** `brand_sentiment` > > Brand sentiment experiments analyze how AI models describe your brand on its own terms. Queries are open-ended questions about your company, product, or category that a real user might ask. > > Results show: > > * An overall sentiment score for each query > * The pros AI models associate with your brand > * The cons AI models associate with your brand > * The specific citations used to support those claims > > Use brand sentiment to understand your baseline AI reputation and identify specific negatives you can address through content and messaging improvements. > [accordion] Competitor comparison > > **Subtype:** `competitor_comparison` > > Competitor comparison experiments analyze how AI models position your brand relative to named competitors. Queries are direct comparison questions ("how does \[your brand] compare to \[competitor]?" or "\[competitor] vs \[your brand]"). > > Results show: > > * A sentiment score reflecting how favorably you're positioned in the comparison > * The pros and cons AI models surface for your brand in the comparison context > * Which competitor brands appear in each response > * Citations supporting the comparison claims > > Use competitor comparison to understand where AI models see you as stronger or weaker than specific competitors, and to identify positioning gaps you can close. ## How sentiment is measured For each query-model combination, XLR8.ai extracts and analyzes the full AI response. The analysis produces: **Sentiment score** — a numeric value on a −5 to +5 scale reflecting the overall tone of the AI's response toward your brand. A higher score indicates more favorable sentiment; a lower score indicates more negative or critical framing. Each score is labeled as one of three polarities, matching what you see on experiment results and Analytics: | Label | Score range | | - | - | | **Positive** | 3.0 to 5.0 | | **Neutral** | 0.0 up to, but not including, 3.0 | | **Negative** | below 0.0 | Ratings and mix percentages (Positive / Neutral / Negative) use these labels. A score such as +1.5 is Neutral, not Positive. **Pros** — the positive attributes, features, or qualities the AI response attributes to your brand. These are extracted as discrete points, each mapped back to specific citations where possible. **Cons** — the negative attributes, drawbacks, or criticisms the AI response attributes to your brand. Like pros, these are extracted as discrete points with citation mappings. On **Detailed Analysis**, **Pros** and **Cons** default to your company name — only phrases attributed to your brand (including company aliases). Switch to **Competitor** to see phrases attributed to other brands, each labeled with the brand name. Phrases without a brand attribution are hidden in both views. **Competitor brands** — for competitor comparison experiments, the specific competitor names that appear in the response alongside your brand. > [!NOTE] > > Sentiment scoring reflects how AI models currently describe your brand based on the sources they have access to. The most effective way to shift sentiment is to create high-quality, citable content that reflects your brand's actual strengths. ## Understanding action item sync For sentiment experiments, XLR8.ai performs an additional processing step after each execution completes: **action item sync**. This step maps the pros, cons, and citations extracted from AI responses to concrete, prioritized action items in the [Action Center](/intelligence/action-center). Each execution has a sync status that tells you whether this processing has completed: | Status | Meaning | | - | - | | `not_started` | The execution completed but action item mapping hasn't begun yet | | `in_progress` | Action item mapping is currently running | | `completed` | Mapping is done — action items are available in the Action Center | **When you see "needs sync":** If an execution shows that action items are not yet synced, you can trigger the sync manually from the experiment detail view. Click **Sync Action Items** on the relevant execution. The sync runs in the background and updates the status automatically. > [!TIP] > > If you've updated your company aliases after running an execution, re-syncing that execution will recompute action items using the updated alias list for more accurate brand mapping. > [!WARNING] > > Only one sync can run at a time per execution. If a sync is already in progress, wait for it to complete before triggering another. ## Interpreting sentiment results When reviewing a completed sentiment execution: **Overall sentiment score** gives you a quick read on whether AI responses about your brand lean positive or negative. Use [Analytics](/analytics/overview) to track sentiment direction across executions. **Pros and cons** are the most actionable outputs. Start with your company name to focus on your brand, then switch to **Competitor** when you want to see how models talk about others. Review the cons list to identify specific criticisms that AI models consistently surface — these represent content and messaging gaps you can address directly. To see how similar phrases group together, open **Reporting → Analytics** and use the **Themes** tab. The theme cluster map starts in 3D so you can rotate clusters; switch to 2D to pan and zoom. Filter the cluster map by aspect, modifier, and intensity for the current Pros or Cons polarity. **Citations** tell you which sources the AI model used to form its response. If the sources cited are competitor content, third-party reviews, or outdated pages about your brand, you have a clear signal about where to focus your content strategy. When tracked URLs are configured, **Detailed Analysis** also shows **Tracked URLs sentiment**, with change vs **Previous execution** or **First execution**. See [Reading results](/experiments/results). **Competitor brands** in competitor comparison results show you which brands the AI model positions you against — and in what framing. If the model consistently pairs you with a competitor you don't consider a direct rival, this can reveal how AI models categorize your product. > [card] [Reading results](/experiments/results) > > Learn how to view execution results, run AI analysis, and track sentiment trends over time. > [card] [Action center](/intelligence/action-center) > > Review and act on the prioritized action items generated from your sentiment results. # Visibility experiments Source: https://docs.tryxlr8.ai/experiments/visibility Measure how often your brand is mentioned and cited by AI models. Configure subtypes, tracked URLs, company aliases, and geo-targeting. A **visibility experiment** tells you whether AI models are including your brand in their responses. XLR8.ai submits your queries to each selected AI model, collects the responses, and checks whether your brand (and its known aliases) appear in the text — and whether your website is cited as a source. Run visibility experiments to answer questions like: * "When users ask Perplexity for the best tools in our category, do we appear?" * "How does our mention rate compare to our top competitors?" * "Which AI models are citing our website as a source?" * "Are we appearing in Google AI Overviews for our target queries?" ## Subtypes Every visibility experiment has a subtype that shapes the kinds of queries XLR8.ai generates. You choose the subtype — and generate or import the queries — in [Query Generation](/query-generation), then pick that query set when you create the experiment. > [accordion] Discovery > > **Subtype:** `discovery` > > Discovery queries are broad, category-level questions a user might ask when exploring options for the first time. They typically take the form "best \[product type] for \[use case]" or "top \[category] tools." > > Use the discovery subtype to measure your brand's presence at the top of the funnel, before a user has a specific product in mind. Strong discovery visibility means AI models include you when users are forming their initial consideration set. > [accordion] Informational > > **Subtype:** `informational` > > Informational queries target users seeking to understand a product category, feature, or concept. They typically take the form "how does \[category] work" or "what is \[feature]." > > Use the informational subtype to measure whether AI models associate your brand with expertise in your domain. Appearing in informational responses signals that AI models treat your content or website as a credible reference. > [accordion] Competitor alternatives > > **Subtype:** `competitor_alternatives` > > Competitor alternatives queries target users who are actively comparing options or looking for alternatives to a named competitor. They typically take the form "alternatives to \[competitor]" or "tools similar to \[competitor]." > > Use this subtype to measure how well you capture demand that originates from competitor awareness. If a user is already considering a competitor and an AI model doesn't mention you as an alternative, you're missing a high-intent opportunity. ## Supported AI models Visibility experiments can run against the models available in the New Experiment picker. The frontend currently exposes: | Model ID | AI model | | - | - | | `perplexity` | Perplexity | | `gpt-fast-gpt-5.1` | GPT Fast (GPT-5.1) | | `gpt-thinking-gpt-5.1` | GPT Thinking (GPT-5.1) | | `gpt-fast-gpt-5.3-chat-latest` | GPT Fast (GPT-5.3 Chat Latest) | | `gpt-thinking-gpt-5.4` | GPT Thinking (GPT-5.4) | | `grok-4` | Grok 4 | | `grok-4.20` | Grok 4.20 | | `gemini` | Gemini | | `claude-opus-4.8` | Claude Opus 4.8 | | `claude-opus-4.7` | Claude Opus 4.7 | | `claude-sonnet-5` | Claude Sonnet 5 | | `claude-sonnet-4.6` | Claude Sonnet 4.6 | | `claude-sonnet-4.5` | Claude Sonnet 4.5 | | `google_ai_mode` | Google AI Mode | | `microsoft_copilot` | Microsoft Copilot | > [!NOTE] > > The live list is fetched from `/experiments/llm-models` when you create an experiment. ## Company aliases When XLR8.ai scans an AI response for your brand, it checks for your company name **and** any aliases you've defined. This matters because AI models often refer to brands by shortened names, abbreviations, or colloquial versions. For example, if your company is "Acme Corporation," an AI model might write "Acme Corp" or simply "Acme." Without aliases, XLR8.ai would not count those as brand mentions. When you create an experiment from a query set, aliases inferred during domain analysis — plus any you added on the Categories step — are copied onto the experiment. **To review or add aliases:** 1. Open the experiment detail page. 2. Open **Settings** and select the **Aliases** tab. 3. Add each alias and save your changes. Aliases apply to all future executions of the experiment. If you add an alias after running executions, re-run the experiment to see updated mention rates that include the new alias. > [!TIP] > > Add common misspellings, abbreviations, ticker symbols, and product names that AI models are likely to use when referencing your brand. ## Tracked competitors **Tracked competitors** are the brands XLR8.ai watches for mention rate alongside yours. Domain analysis in [Query Generation](/query-generation) infers competitors for every subtype. For **Competitor alternatives**, the names also match the competitor list on that step. Creating an experiment from the query set saves those names on the experiment. You can still add or edit them in **Settings → Competitors**. Saving competitor names recomputes mention rate from existing results — you do not need to re-run the experiment. ## Tracked URLs **Tracked URLs** let you monitor whether specific pages on your website — like your homepage, a landing page, or a product page — are being cited as sources by AI models. When you configure tracked URLs, XLR8.ai identifies which citations in AI responses match your URLs and computes a **tracked URL impact** score that quantifies how much citation traffic those pages are generating across your queries. On **Source Analytics**, compare citation change to **Previous execution** or **First execution**. Those comparisons load after the rest of the results page. See [Reading results](/experiments/results). **To configure tracked URLs:** 1. Open the experiment detail page. 2. In the **Tracked URLs** section (visible only on visibility experiments), click **Edit**. 3. Enter each URL you want to track, one per line. Duplicate and blank entries are automatically removed. 4. Save your changes. You can also add one or more URLs across several experiments at once from **Settings → Tracked URLs**. Paste a list (one per line or comma-separated) instead of editing each experiment individually. See [Settings](/integrations/settings#tracked-urls). > [!NOTE] > > Tracked URLs are only available on visibility experiments. The impact analysis appears alongside your standard result analysis once URLs are configured and at least one execution has completed. ## Geo-targeting Geo-targeting controls the geographic context used when submitting queries to AI models. This is useful when your brand has a regional focus or when competitive dynamics differ significantly by market. You can target at three levels of specificity: | Field | Example | | - | - | | Country | `US`, `GB`, `AU` (ISO alpha-2 codes) | | Region | `California`, `Ontario` | | City | `San Francisco`, `London` | Set geo-targeting when you create the experiment. It applies to all executions of that experiment. > [!TIP] > > If you sell globally, consider creating separate experiments for your key markets rather than targeting a single location. This gives you clean, comparable data per region. ## How visibility scores are calculated After each execution, XLR8.ai analyzes the collected AI responses and computes visibility metrics for your brand across every query and model combination. **Mention rate** is the primary metric. It measures what percentage of queries resulted in your brand being mentioned in the AI response. ```text theme={null} Mention rate = (queries where brand was mentioned) / (total queries) × 100 ``` Beyond the overall mention rate, results are broken down by: * **Model** — how your mention rate differs across AI models * **Category** — which query categories are driving the most visibility * **Individual query** — the raw AI response and whether your brand appeared **Citation presence** is tracked separately. A citation means your website URL appeared in the list of sources that the AI model used or referenced in its response. High citation presence indicates that AI models are actively using your content as a reference. **Competitor brands** are extracted from the same AI responses. For each query where competitors appear alongside (or instead of) you, XLR8.ai records which brands were mentioned. This lets you see your share of voice relative to the competitive landscape. > [card] [Reading results](/experiments/results) > > Learn how to interpret visibility results and citation data. > [card] [Analytics](/analytics/overview) > > Track how visibility changes across executions. > [card] [Outreach](/outreach/overview) > > Turn visibility citations into listicle outreach campaigns. > [card] [Scheduling](/experiments/scheduling) > > Set up automatic recurring executions to track how your visibility changes over time. # Indexing Source: https://docs.tryxlr8.ai/indexing/overview Queue URLs for Google Search Console and Bing Webmaster indexing, and monitor submission status. Indexing lets you submit URLs to **Google Search Console** and **Bing Webmaster** from XLR8.ai. Properties are matched per URL automatically. Navigate to **General → Indexing** in the sidebar, or open **Configure Indexing** from Settings → Integrations. *** ## Connect Google and Bing On the Indexing page: 1. Click **Connect** for **Google Search Console** and/or **Bing Webmaster** 2. Complete OAuth for the account that owns your properties 3. Confirm the connected badge shows your site count If a connection expires, that engine shows **Reconnect required** instead of **Not connected**, with a note that the token has expired. Click **Reconnect** and complete OAuth again to restore access. To disconnect an engine, use **Disconnect**. Disconnecting clears queued URLs for that engine; completed history is preserved. *** ## Submit URLs to the queue **1. Choose engines** Toggle **Google** and/or **Bing**. Engines that are not connected appear disabled. **2. Add URLs** Paste full `https://` URLs into the textarea, or use **Upload CSV** / **Sample CSV**. **3. Submit** Click **Submit to queue**. Submissions are processed by a drain job. Indexing status updates hourly. *** ## Queue statuses Filter the queue by status and date range: * Queued * Submitted * Last discovered * Indexed * Failed Each row shows the URL, matched property, and per-engine progress for Google and Bing. If you hit a daily quota for an engine, XLR8.ai surfaces that error in the UI. When Google Search Console cannot verify site ownership or inspect permissions for a URL, a **Verify** error appears under that engine's status with the reason. *** ## Content staleness Under **General → Indexing → Content staleness**, monitor blog URLs for freshness. Add blog URLs so XLR8.ai can flag content that may need a refresh before you re-optimize or re-index it. *** ## Related > [card] [Visibility experiments](/experiments/visibility) > > Track whether indexed pages start appearing as citations. > [card] [Settings](/integrations/settings) > > Open indexing and other integrations from Settings. > [card] [Content optimization](/content/optimization) > > Refresh stale pages before submitting them for indexing. > [card] [Reporting](/reporting/overview) > > Review Search Console and Analytics performance after pages are live. # Settings Source: https://docs.tryxlr8.ai/integrations/settings Configure your company profile, blog URL monitoring, and third-party integrations from a single place. The **Settings** page is where you configure the details XLR8.ai uses across every feature — your company identity, brand guidelines, schedules, navigation preferences, `llms.txt`, and third-party integrations. Navigate to **Settings** from the **General** section of the main sidebar at any time. Invoices and payment methods live under **Settings → Billing** — see [Billing](/account/billing). *** ## Company profile Under **Company Profile → Company Name, Domain & Internal Team**, set: * **Company Name** — your brand name as it should appear in experiments, content, and Prospect Intelligence * **Company Domain** — your primary domain without a protocol (for example, `acme.com`). Prospect Intelligence uses it the same way to recognize your team in transcripts * **Internal Team** — comma-separated names of people on your side of a call. [Prospect Intelligence](/intelligence/prospect-intelligence) uses these to keep prospect turns and skip your team's speech * **Third-Party Reviews Domain** (optional) — used when analyzing third-party review sources XLR8.ai pre-fills your company name and domain when you create experiments, generate content, and configure social listeners. *** ## Organization brand guidelines Under **Brand Guidelines → Organization Brand Guidelines**, admins can provide: * Freeform guidance text * PDF uploads (max 10MB each) * A toggle to **Default to including brand guidelines in content tools** Members can view guidelines; only organization admins can edit them. When the default toggle is on, content generation and related tools include these guidelines automatically. If an uploaded PDF is a link-placement brief, it also appears in the **Link plans** list on the same card (plan name, rule count, and source filename). On a generated content piece, click **Apply link plan** to insert those placements and review them before accepting. *** ## AI Brand Learnings Under **Brand Corrections → AI Brand Learnings**, review corrections captured from content comment threads. On a generated content piece, check threads in the **Comments** sidebar and click **Remember** to send them here. Anyone in the organization can: * Filter by **pending**, **approved**, **dismissed**, or **all** * Edit a learning's text (except dismissed items) * **Apply to future content** to approve a pending learning Organization admins can also: * **Add correction** manually * **Import from doc** * **Dismiss** a pending learning * Delete a learning Approved learnings help keep future generated content consistent with your brand. *** ## Experiment schedules Under **Experiment Schedules → Schedule Overview**, organization admins can view and manage automated run schedules across experiments. See [Scheduling](/experiments/scheduling) for cadence options and how schedules work. *** ## Tracked URLs Under **Tracked URLs → Add URLs to multiple experiments**, add the same pages to several visibility or sentiment experiments at once. 1. Paste one or more URLs into **URLs** — one per line, or separated by commas 2. Optionally choose a **bucket** for analytics grouping. A chosen bucket applies to every pasted URL. If you leave it empty, XLR8.ai fills a bucket per URL from the site (for example Reddit, YouTube, LinkedIn, or PDF) 3. Select the experiments to update 4. Click **Add to selected experiments** Only active visibility and sentiment experiments appear in the list. Duplicates are skipped per experiment. After you submit, results show which experiments were updated, already tracking the URLs, or failed. To manage tracked URLs on a single experiment, use that experiment's **Settings → Tracked URLs** tab. See [Visibility experiments](/experiments/visibility#tracked-urls). *** ## Navigation — Additional Tools Under **Navigation → Additional Tools**, toggle **Show Additional Tools in the sidebar**. When enabled, the sidebar shows: * **Additional Tools** — Content Alignment and Cosine Similarity Maximiser * **AEO Audit** — homepage AI discoverability scorecard This preference is stored in your browser (localStorage) and defaults to off. *** ## LLM.txt Under **AI Models & Content Settings → LLM.txt**, generate a downloadable `llms.txt` file for your site. Provide: * **Company name** * **Company URL** Then click **Generate & download**. *** ## Content staleness Blog URL monitoring lives under **General → Indexing → Content staleness** (`/settings/content-staleness`), not on the main Settings page. Add blog URLs to monitor content freshness. See [Indexing](/indexing/overview) for related URL submission workflows. *** ## Integrations The **Integrations** section links to configuration pages for third-party tools: > [card] [Webflow](/integrations/webflow) > > Publish generated content to Webflow CMS collections. > [card] [WordPress](/integrations/wordpress) > > Connect WordPress sites over XML-RPC and publish posts or pages. > [card] [Indexing](/indexing/overview) > > Connect Google Search Console and Bing Webmaster, then queue URLs for indexing. > [card] [Reporting](/reporting/overview) > > Connect Google Search Console and Analytics, and upload monthly attribution CSVs. > [card] [Outreach Email](/outreach/overview) > > Connect Gmail, set a sender name and client tagline, and send listicle outreach from XLR8.ai. *** ## Related > [card] [Content generation](/content/generation) > > Generate AI-optimized content using your company profile and brand guidelines. > [card] [Prospect Intelligence](/intelligence/prospect-intelligence) > > Upload call transcripts using the company name, domain, and Internal Team from Settings. > [card] [Experiments](/experiments/overview) > > Run visibility and sentiment experiments with your company domain. > [card] [Authentication](/account/authentication) > > Manage organizations, roles, and team access. > [card] [Billing](/account/billing) > > View invoices, pay on Stripe or in XLR8, and manage your payment method. > [card] [Quick start](/quickstart) > > New to XLR8.ai? Start here. # Webflow Integration Source: https://docs.tryxlr8.ai/integrations/webflow Publish XLR8.ai-generated content directly to your Webflow CMS — no copy-pasting required. The Webflow integration lets you push content generated by XLR8.ai straight into your Webflow site as CMS collection items. Once connected, you can select a target site and collection, then publish in a single action from the XLR8.ai content view. ## Prerequisites Before connecting, make sure you have: * A Webflow account with at least one published site * A CMS collection on that site (or you can create one during setup) * A Webflow **personal access token** with the following scopes: `sites:read`, `sites:write`, `cms:read`, `cms:write` > [!NOTE] > > You need a site-level token, not a workspace token. Site tokens are created under **Site Settings → Apps & integrations → API access** in the Webflow dashboard. *** ## Connect your Webflow account **1. Open the Webflow integration page** In XLR8.ai, go to **Settings → Integrations → Webflow** (or navigate directly to `/settings/integrations/webflow`). **2. Generate a Webflow personal access token** In a separate tab, open your Webflow site dashboard and go to **Site Settings → Apps & integrations → API access**. Click **Generate API token**, give it a name such as "XLR8 Integration", and grant **Read and Write** access for all four scopes: * `sites:read` * `sites:write` * `cms:read` * `cms:write` Click **Generate token** and copy it. You won't be able to see it again. > [!TIP] > > Click **How to get a token?** on the XLR8.ai integration page for a step-by-step guide with a link to the official Webflow documentation. **3. Paste the token and connect** Back in XLR8.ai, paste the token into the **Webflow personal access token** field and click **Connect with API token**. XLR8.ai validates the token by calling the Webflow API and retrieving your sites. If the token is valid, the connection status changes to **Connected** and your first site is selected automatically. > [!WARNING] > > Your token is stored encrypted and used only to call the Webflow API on your behalf. Never share your token publicly or include it in version-controlled files. *** ## Configure your site and collection After connecting, the integration page shows a **Connected** status with your Webflow site name. Configure where your content will be published before you use the publish action. ### Select a site Use the **Webflow Site** dropdown to choose which of your Webflow sites receives published content. Switching the site reloads the available collections automatically. ### Select a default collection Use the **Default Collection** dropdown to choose the CMS collection that XLR8.ai will publish content into. This collection is used by default every time you publish a piece of content. If you don't have a suitable collection yet, click **Create Collection** to create one directly from XLR8.ai. #### Create a collection **1. Click Create Collection** Select the target site first, then click **Create Collection** next to the collection dropdown. **2. Enter a name and optional slug** Enter a base name for the collection — XLR8.ai automatically prefixes it with `XLR8-` to keep it identifiable in your Webflow dashboard (for example, `XLR8-Blog Posts`). Optionally enter a URL slug. If you leave the slug empty, it is generated from the name. The collection is created with **Name** and **Body** fields. **3. Confirm** Click **Create Collection**. XLR8.ai creates the collection in Webflow and selects it as the default automatically. > [!NOTE] > > Collections created from XLR8.ai include a Name field and a Body field. If you need additional fields (such as author, category, or tags), add them in the Webflow CMS editor after creation. *** ## Publish content to Webflow Once your site and collection are configured, you can publish any piece of generated content to Webflow. **1. Open a content item** Navigate to **Content** and open the piece of content you want to publish. **2. Select Publish to Webflow** Click the **Publish to Webflow** action. XLR8.ai uses your configured default collection. **3. Review the slug** A URL slug is generated from your content title. You can customize it before confirming. Slugs must be lowercase, hyphen-separated, and unique within the collection. > [!TIP] > > XLR8.ai checks slug availability in real time. If a slug is already taken, it suggests an alternative. **4. Choose draft or live, then confirm** **Publish immediately** is off by default. Leave it off and click **Save as Draft** to create a Webflow CMS item as a draft. Turn the toggle on and click **Publish** to make the item live. After you save a draft, the content toolbar shows a draft badge and a **Publish Draft** action. That dialog defaults to publishing immediately so you can go live when you are ready. *** ## Check sync status After publishing, XLR8.ai tracks whether your local content has changed since the last sync to Webflow. If you edit the content in XLR8.ai after publishing, the content view shows a **local changes** indicator. This means your Webflow CMS item is out of date relative to the current version in XLR8.ai. To push the latest version to Webflow, use the **Update** (or re-publish) action. XLR8.ai sends the updated fields and marks the content as synced. Updates keep the item in its current Webflow state — a draft stays a draft, and a live item stays live. *** ## Disconnect Webflow To remove the Webflow integration from your account: 1. Go to **Settings → Integrations → Webflow**. 2. Click **Disconnect**. 3. Confirm the prompt. Disconnecting removes the stored access token and integration record. Previously published content remains in your Webflow site — disconnecting does not delete any Webflow CMS items. > [!WARNING] > > After disconnecting, you won't be able to publish new content or update existing Webflow items from XLR8.ai until you reconnect. *** ## Troubleshooting > [accordion] Connection fails with 'Invalid token or missing required scopes' > > Your token is missing one or more required scopes. Go to **Webflow → Site Settings → Apps & integrations → API access**, revoke the existing token, generate a new one, and make sure you enable **Read** and **Write** for `sites:read`, `sites:write`, `cms:read`, and `cms:write`. > [accordion] No sites appear after connecting > > Your token must be a **site-level token**, not a workspace-level token. Site-level tokens are generated from within a specific site's settings, not from the Webflow workspace dashboard. Create a new token from **Site Settings → Apps & integrations → API access**. > [accordion] No collections appear for my site > > The selected site may not have any CMS collections. In Webflow, go to **CMS → Collections** for that site and create at least one collection. Then return to XLR8.ai and reload the collections list. Alternatively, use **Create Collection** in XLR8.ai to create one directly. > [accordion] Slug validation error when publishing > > Slugs must contain only lowercase letters, numbers, and hyphens. They cannot start or end with a hyphen, and must be unique within the collection. Edit the slug to remove any special characters or spaces, and try again. > [accordion] Content published but changes aren't visible on the live site > > Webflow may require a site publish after CMS item changes. Go to your Webflow site dashboard and click **Publish** to push the changes live. If the issue persists, check that your token has `sites:write` scope — without it, automatic site publishing after content changes is not possible. > [accordion] Disconnect succeeded but the item is still in Webflow > > Disconnecting XLR8.ai only removes the integration credentials — it does not delete CMS items from Webflow. To remove published items, delete them directly from the Webflow CMS editor. *** ## Related > [card] [Content generation](/content/generation) > > Generate AI-optimized content ready to publish to Webflow. > [card] [Content optimization](/content/optimization) > > Improve existing content before pushing it to your Webflow site. > [card] [Settings](/integrations/settings) > > Configure your company profile and blog URL monitoring. > [card] [Quick start](/quickstart) > > New to XLR8.ai? Start here. # WordPress Integration Source: https://docs.tryxlr8.ai/integrations/wordpress Connect WordPress sites over XML-RPC and publish XLR8.ai-generated content as posts or pages. Publish XLR8.ai-generated content directly to WordPress over XML-RPC. You can connect multiple sites, verify credentials, and publish as a live post/page or as a draft. *** ## Connect a WordPress site **1. Open WordPress settings** Go to **Settings → Integrations → WordPress**, or open `/settings/integrations/wordpress`. **2. Add a site** Click **Add site** and fill in: * **Label** * **XML-RPC URL** — typically `https://YOUR_SITE_URL/xmlrpc.php` * **Username** * **Password / app password** * **Public site URL** (optional) **3. Verify credentials** Click **Verify credentials** on the site card to confirm the connection works. You can connect multiple WordPress sites and mark them Active or Inactive. *** ## Publish from content generation **1. Open a generated content piece** Go to **Content Generation** and open the item you want to publish. **2. Click Publish to WordPress** Use the publish toolbar action **Publish to WordPress**. **3. Configure the publish dialog** Choose: * **Site** * **Content type** — Blog post or Page * **Parent page** (for pages) * **URL slug** * **Publish immediately** — off by default; leave it off to save as a WordPress draft, or turn it on to publish live **4. Publish** Click **Publish**. With the toggle off, WordPress receives a draft and the toolbar shows **WordPress · Draft**. Turn **Publish immediately** on to go live (**WordPress · Live**). If the item is already a draft on WordPress, use **Publish Draft** on the toolbar. That dialog defaults to publishing immediately. If no sites are configured, the dialog links you to **Open WordPress settings**. *** ## Related > [card] [Content generation](/content/generation) > > Generate the content you want to publish. > [card] [Webflow integration](/integrations/webflow) > > Publish to Webflow CMS instead of WordPress. > [card] [Settings](/integrations/settings) > > Manage all workspace integrations from one place. # Action Center Source: https://docs.tryxlr8.ai/intelligence/action-center Track and act on prioritized recommendations generated from your experiment results. The Action Center is a unified inbox for recommendations derived from your XLR8.ai experiments. Each time an experiment execution completes, the platform analyzes the results — visibility gaps, sentiment patterns, low-performing query categories — and surfaces them as actionable items so you always know what to work on next. Action items are organized, filterable, and trackable. You can mark them complete as you address each recommendation, and reopen them if circumstances change. ## Accessing the Action Center Click **Action Center** in the left sidebar. The page loads with all pending action items from your experiments, sorted by most recent by default. ## Understanding action items Each action item represents a specific finding from an experiment result. The item tells you: * **Source** — which experiment (or other source) generated the recommendation * **Category** — the query category or topic area the finding relates to * **Action data** — the specific insight, such as a list of gaps or negative sentiment themes * **Created date** — when the finding was detected * **Status** — whether the item is pending or has been marked complete ### Action source types > [tab] Experiment > > Items sourced from **experiment** results are the most common. They are generated when an execution completes and the platform identifies: > > * Query categories with low or zero visibility > * Competitors outperforming you across specific models > * Negative sentiment trends requiring content updates > * Significant drops in mention rate compared to previous executions > > Each experiment-sourced item links back to the originating experiment so you can review the underlying data. > [tab] Content > > Items sourced from **content** analysis flag opportunities or issues related to your published content. These appear when content optimization workflows detect recommendations that require your attention. ## Filtering and sorting Use the filter bar at the top of the Action Center to narrow down the list to what matters most right now. | Filter | Description | | - | - | | **Source** | Show items from a specific source type (`experiment` or `content`). | | **Experiment name** | Search for action items from a particular experiment by name. | | **Category** | Filter by query category (e.g., "Healthcare", "Pricing", "Competitor Comparison"). | | **Date range** | Show only items created within a specific date window. | | **Status** | Toggle between **Pending** (open items) and **Completed** (resolved items). | | **Tags** | Filter by one or more tags attached to action items. | You can combine multiple filters at the same time. To reset, clear all active filters. > [!TIP] > > Start each week by filtering to **Pending** items sorted by **Newest first**. This surfaces the freshest experiment findings before older items accumulate. ### Sorting Click the **Sort** control to change the ordering. Available sort options: * **Newest first** (default) — most recently created items appear at the top * **Oldest first** — useful for working through a backlog in order ## Marking items complete When you have addressed a recommendation — for example, updated your content to close a visibility gap — mark the action item as complete to keep your list clean. **1. Find the action item** Locate the item in the list. Use filters to narrow down to the relevant experiment or category if needed. **2. Mark as complete** Click the **Mark complete** button or checkbox on the action item card. The item moves to the **Completed** view immediately. **3. Confirm in Completed view** Switch the status filter to **Completed** to verify the item was recorded, along with the timestamp of when it was completed. > [!NOTE] > > Completion is recorded per organization. Any member of your organization can mark items complete, and the record shows who completed each item and when. ## Reopening a completed item If you need to revisit a recommendation — for example, if a follow-up execution shows the issue persists — you can reopen a completed item. **1. Switch to Completed view** Set the **Status** filter to **Completed** to see your resolved items. **2. Reopen the item** Click **Reopen** on the completed item card. The item returns to **Pending** status and reappears in your active list. ## Working through your backlog > [!TIP] > > Focus on action items tied to your highest-traffic query categories first. Improvements to high-volume categories have the largest impact on overall AI visibility. > [!TIP] > > Use the **Experiment name** filter to review all recommendations from a single experiment at once. This is useful after a major execution completes and you want a comprehensive view of findings. > [!WARNING] > > Action items are generated automatically from experiment results. New items appear after each execution completes. Check the Action Center regularly after scheduling automated runs to stay on top of fresh recommendations. ## Related > [card] [Experiments overview](/experiments/overview) > > Learn how experiments generate the data behind action items. > [card] [Experiment results](/experiments/results) > > Review the raw results that action items are derived from. > [card] [AI Insights Agent](/intelligence/insights-agent) > > Ask questions about your experiment data in plain language. > [card] [Scheduling](/experiments/scheduling) > > Automate executions so action items stay current. # AEO Audit Source: https://docs.tryxlr8.ai/intelligence/aeo-audit Run a homepage-only AEO audit to score AI discoverability and get fix suggestions. AEO Audit runs a homepage-only scorecard for AI discoverability. It evaluates your homepage and returns an overall AEO score, category scores, and check-level recommendations. *** ## In-app audit In the authenticated app: 1. Enable **Show Additional Tools in the sidebar** under **Settings → Navigation** 2. Open **Execution → AEO Audit** 3. Enter your **Homepage URL** 4. Click **Run audit** The report shows: * **Overall AEO Score** * Category cards labeled **Good**, **Average**, or **Needs work** * Per-check scores with optional **fix** guidance *** ## Public audit A public lead-capture version is available at `/aeo-audit` without signing in. That form asks for: * **Full name** * **Email** * Homepage URL After the audit completes, XLR8.ai can email an AI Visibility Report within 24 hours. *** ## When to use it Use AEO Audit when you want a fast homepage health check before or alongside a full experiment. It does not replace visibility or sentiment experiments — it complements them with site-level discoverability guidance. *** ## Related > [card] [Experiments](/experiments/overview) > > Measure brand mentions and sentiment across AI models. > [card] [Settings](/integrations/settings) > > Enable Additional Tools so AEO Audit appears in the sidebar. > [card] [Content optimization](/content/optimization) > > Improve existing pages based on audit and experiment findings. # AI Insights Agent Source: https://docs.tryxlr8.ai/intelligence/insights-agent Ask natural language questions about your experiment data and get instant, AI-powered answers about your AEO performance. The AI Insights Agent is a conversational assistant built directly into your experiment results. Instead of manually scanning tables and charts, you can ask plain-language questions and receive data-driven answers grounded in your specific experiment and execution results. The agent is context-aware: it knows which experiment and execution you are viewing, so your questions do not need to include IDs or technical identifiers. ## Accessing the Insights Agent Open any experiment from the **Experiments** dashboard, then navigate to its results view. The Insights Agent chat panel appears alongside your results data. You can open a conversation at any time while reviewing an execution. > [!NOTE] > > The agent operates within the scope of the experiment and execution you are currently viewing. It does not search the general web or access data outside your XLR8.ai account. ## Asking questions Type your question in the chat input and press **Enter** or click **Send**. The agent responds with an analysis grounded in your experiment data. You can ask follow-up questions to dig deeper — the agent maintains context across the conversation. ### Example questions The following questions reflect common AEO analysis workflows. Use them as a starting point or inspiration for your own queries. > [accordion] Model coverage and mention rates > > * "Which model mentions us most frequently?" > * "Which models never cite us in this execution?" > * "How does our mention rate compare across GPT-4o, Perplexity, and Gemini?" > * "Which query categories have the lowest mention rate?" > [accordion] Visibility trends over time > > * "Compare our visibility across the last 3 executions." > * "Has our mention rate improved since we updated our content?" > * "Show me which queries saw the biggest change between execution 1 and execution 3." > [accordion] Competitive analysis > > * "How does our visibility compare to Competitor X in this execution?" > * "Which models favor our competitors over us?" > * "Where are we beating Competitor X, and where are we losing?" > [accordion] Citation and gap analysis > > * "What are the main reasons we're not being cited?" > * "Which queries consistently fail to mention us?" > * "Are there query categories where we have zero visibility?" > [accordion] Sentiment and tone > > * "What is the overall sentiment toward our brand in this execution?" > * "Which models describe us most positively?" > * "Are there any negative themes in how we are being mentioned?" ## Conversation flow The agent maintains the full history of your conversation within the current session. You can build on previous answers to drill down into specific findings. **1. Start with a broad question** Open with a high-level question, such as "Summarize our visibility results for this execution." **2. Follow up on specifics** Ask follow-up questions to investigate particular models, categories, or competitors: "Why is our mention rate low on Perplexity specifically?" **3. Request actionable takeaways** Close the conversation with an action-oriented question: "What should we focus on to improve our visibility in the next execution?" ## Tips for useful answers > [!TIP] > > Be specific about what you want to compare. Questions like "Compare our visibility on GPT-4o vs. Gemini for the Healthcare category" produce more focused answers than "How are we doing?" > [!TIP] > > If you are viewing a specific execution, the agent focuses on that execution's data. Switch to a different execution in the results view before asking execution-specific questions. > [!TIP] > > Ask the agent to summarize before you dig into details. A summary response often reveals the most important finding and guides your next question. ## Limitations > [!WARNING] > > The Insights Agent answers questions based on your experiment data only. It cannot browse the web, fetch live search results, or access data from outside your XLR8.ai account. * Conversations are session-scoped. Starting a new session clears the conversation history. * The agent focuses on the experiment and execution you are currently viewing. If you need to compare across multiple experiments, navigate to each one separately. * Answers reflect the data captured at the time the execution ran. The agent does not have access to real-time AI model outputs. ## Related > [card] [Experiments overview](/experiments/overview) > > Learn how experiments are structured and how executions are run. > [card] [Visibility experiments](/experiments/visibility) > > Understand how visibility data is collected across AI models. > [card] [Sentiment experiments](/experiments/sentiment) > > See how sentiment is measured and scored per execution. > [card] [Action Center](/intelligence/action-center) > > Turn insights into prioritized recommendations. # Prospect Intelligence Source: https://docs.tryxlr8.ai/intelligence/prospect-intelligence Upload prospect call transcripts to see what they talked about, asked, praised, and objected to. Prospect Intelligence analyzes prospect call transcripts so you can see what they talked about, asked, praised, and objected to. XLR8.ai extracts statements, groups them by theme and type, and lets you open the supporting quotes. ## Accessing Prospect Intelligence From the sidebar under **Strategy**, click **Prospect Intelligence**. ## Set company identity first Company name, domain, and Internal Team come from [Settings](/integrations/settings). Prospect Intelligence uses them to tell your brand from the prospect and to keep prospect turns while skipping your team's speech. **1. Open Settings** Go to **Settings** and open the company profile card (**Company Name, Domain & Internal Team**). **2. Save company name and domain** Enter your **Company Name** and **Company Domain**. These identify your brand in the transcript. **3. Add Internal Team names** In **Internal Team**, enter comma-separated names of people on your side of the call. Prospect Intelligence uses these so it can skip your team's speech and keep the prospect's turns. > [!TIP] > > Update Internal Team when new sellers join calls. Missing names can mix your team's speech into the prospect analysis. ## Upload transcripts **1. Choose files** Under **Transcripts or Zip**, select one or more files. Accepted formats include `.txt`, `.vtt`, `.srt`, `.docx`, `.pdf`, `.md`, and `.zip`. **2. Upload** Click **Upload**. Files are grouped into a batch. Status moves from **Pending** to **Processing**, then **Succeeded** or **Failed**. **3. Refresh if needed** The page polls automatically while a batch is processing. Click **Refresh** to reload immediately. > [!NOTE] > > Processing can take a few minutes for large batches. Failed files show an error message you can review before you delete and retry. ## Previously Added Uploaded batches appear under **Previously Added**. Expand a batch to see each file, its status, and how many statements were extracted. You can delete an entire batch or a single transcript. Deleting removes that analysis from the charts and review list. ## Filter by theme and type Use the filters above the charts to focus the view: | Filter | Description | | - | - | | **Themes** | Limit results to selected themes such as Pricing & budget, ROI & value, Product & features, Competition, or Next steps. Leave empty for **All Themes**. | | **Type** | Limit results to **Question**, **Pain**, **Objection**, **Request**, **Praise**, or **Statement**. **All** shows every type. | Theme filters apply to the charts, heatmap, and **Calls To Review**. Type filters apply to the charts. ## Theme charts After extraction succeeds, charts appear only when they have data: * **Most Talked About** — themes that appear in the most calls * **Most Asked** — themes with the most questions * **Most Praised** — themes with the most praise * **Most Painful** — themes with the most pain * **Most Objected** — themes with the most objections * **Most Requested** — themes with the most requests Each bar is a count of unique calls. Click a bar to open the statements and quotes for that theme. ## Theme By Type The **Theme By Type** heatmap shows how often each theme appears as each type. Switch: * **By topic** — how many times that theme and type appear * **By call** — how many unique calls include that theme and type Click a cell to open the matching statements. In **By call** mode, quotes are grouped by call. ## Calls To Review **Calls To Review** lists each call with counts per type. It uses the theme filter above. Choose which type columns to show — **Total** is only those types. Click a column header to sort, and click it again to reverse direction. Click a row to open that call's statements and quotes. ## Reading quotes The quote drawer shows the extracted point, themes, type, and supporting quotes. Speaker and company appear only when they were in the transcript. ## Use themes in Query Generation After calls are analyzed, [Query Generation](/query-generation) can cite those structured themes when it generates or tags queries. On the Company step, a note under Additional context appears when this organization has Prospect Intelligence items. Click a **Sales call** citation badge on a query to see the theme, type, and transcript. ## Related > [card] [Query Generation](/query-generation) > > Build query sets that can cite Prospect Intelligence themes. > [card] [Settings](/integrations/settings) > > Save company name, domain, and Internal Team names used on transcripts. > [card] [Action Center](/intelligence/action-center) > > Track prioritized recommendations from experiment results. > [card] [AI Insights Agent](/intelligence/insights-agent) > > Ask questions about your experiment data in plain language. > [card] [Quick start](/quickstart) > > Set up your organization and run your first experiment. # Social Listening Source: https://docs.tryxlr8.ai/intelligence/social-listening Monitor Reddit and X (Twitter) for brand mentions, category discussions, and content opportunities relevant to your business. Social Listening lets you track what real users are saying about your brand, your category, and your competitors across Reddit and X (Twitter). Instead of manually searching social platforms, XLR8.ai surfaces relevant posts, scores them by opportunity type, and lets you save the most valuable ones for future reference or content generation. Use Social Listening to: * Discover how users describe problems your product solves * Identify content gaps and questions you could answer * Track competitor mentions and sentiment across communities * Generate AI-powered content responses grounded in real discussions > [tab] Reddit > > ## Reddit Intelligence > > Reddit is organized into communities (subreddits) where users ask questions, share experiences, and debate topics in depth. XLR8.ai analyzes posts across the subreddits most relevant to your business and surfaces those with the highest engagement opportunity. > > ### Setting up your Reddit listener > > Before running an analysis, configure your listener profile so XLR8.ai knows what to look for. > > **1. Open Social Listening** > > Click **Social Listening** in the left sidebar. The Reddit tab opens by default. > > **2. Run auto-setup** > > Click **Set up listener**. XLR8.ai reads your existing experiment data to generate a company description and suggest relevant subreddits automatically. > > > [!NOTE] > > > > If you have not run any experiments yet, you will be prompted to enter your company name and domain manually so the platform can generate your profile. > > **3. Review your company description** > > Check the generated company description. This text is used to determine what counts as a relevant post. Edit it to better reflect your product, use case, or target audience. > > **4. Confirm your subreddits** > > Review the suggested subreddits. Add or remove communities to match where your audience actually participates. Enter subreddit names without the `r/` prefix. > > **5. Save your profile** > > Click **Save profile**. Your listener configuration is stored and used for all future Reddit analyses. > > ### Running a Reddit analysis > > Once your profile is set up, run an analysis to find relevant posts. > > **1. Set your parameters** > > Choose how many days back to scan (up to 30 days) and the maximum number of posts to return (up to 100). A **Days back** value of 7 with **Top posts** set to 20–30 is a good starting point for weekly monitoring. > > **2. Start the analysis** > > Click **Find relevant posts**. The analysis runs in the background — you can continue using the platform while it processes. The status updates automatically when it completes. > > **3. Review results** > > Posts are returned ranked by relevance score (0–10). Each post shows: > > * The post title and subreddit > > * Upvote count and comment count > > * **Opportunity type** — one of: Answer Question, Share Expertise, Provide Solution, Join Discussion > > * **Reasoning** — a brief explanation of why the post is relevant and how your brand could add value > > > [!TIP] > > > > Re-run the analysis weekly to stay current. Reddit conversations move quickly, and fresh posts often represent the highest-value engagement opportunities. > > ### Filtering results > > Use the filter controls above the post list to refine what you see: > > * **Days back** — adjust the look-back window to see older or newer posts > * **Number of posts** — increase the limit if you want broader coverage > > ### Saving posts > > Save posts you want to revisit, use for content inspiration, or pass to your team. > > **1. Find a post to save** > > Browse the analysis results and identify a post worth saving. > > **2. Save the post** > > Click the **Save** button on the post card. The post is added to your **Saved posts** tab. > > **3. Access saved posts** > > Switch to the **Saved** tab to see all posts you have bookmarked. Saved posts persist across sessions and analyses. > > ### Generating AI-powered content from Reddit posts > > For any post you find relevant, you can generate a suggested comment or response grounded in the discussion. > > **1. Open a relevant post** > > Find a post in your results or saved list that you want to respond to. > > **2. Generate content** > > Click **Generate content**. XLR8.ai analyzes the post and produces a tailored response that positions your brand as a helpful expert — not a promotional spam message. > > **3. Review and refine** > > Read the generated content carefully. Adjust the tone or specifics before posting it on Reddit. Always review AI-generated content before publishing. > > > [!WARNING] > > > > Generated content is a starting point, not a final draft. Reddit communities are sensitive to promotional or inauthentic responses. Personalize and review all generated content before posting. > [tab] X (Twitter) > > ## X Intelligence > > X (Twitter) moves fast and surfaces real-time sentiment, trending topics, and customer conversations. XLR8.ai finds posts relevant to your brand and category, scores them by engagement opportunity, and lets you save the most valuable ones. > > ### Setting up your X listener > > **1. Open X Intelligence** > > In the sidebar under **Execution → Social Listening**, click **X Intelligence**. > > **2. Configure your company description** > > Enter or confirm your company description. This text tells XLR8.ai what topics, keywords, and discussions are relevant to your brand. Click **Save profile** to store it. > > > [!NOTE] > > > > Unlike Reddit, X Intelligence does not require subreddit configuration. Your company description is the primary signal used to find relevant posts. > > ### Finding relevant X posts > > **1. Set your parameters** > > Choose the **Days back** window (up to 30 days) and the number of **Top posts** to return (up to 100). > > **2. Run the search** > > Click **Find relevant posts**. XLR8.ai scans recent X posts and returns those most relevant to your brand and category. > > **3. Review results** > > Each post shows: > > * Post text and author > * Engagement metrics: likes, retweets, replies > * **Opportunity type** — one of: Engage in Discussion, Share Expertise, Provide Solution, Customer Support > * **Reasoning** — why the post is relevant and how your brand could add value > > ### Saving X posts > > **1. Save a post** > > Click **Save** on any post card in the results. The post is added to your saved X posts list. > > **2. View saved posts** > > Switch to the **Saved** tab to browse all your saved X posts. Use this list to track conversations you plan to engage with or reference for content. > > **3. Remove saved posts** > > Select one or more posts in the saved list and click **Delete** to remove them in bulk. ## Use cases > [card] [Content opportunity discovery](/content/generation) > > Find unanswered questions in your category and turn them into content that AI models will cite. > [card] [Competitive monitoring](/experiments/visibility) > > Spot conversations where competitors are being recommended instead of you. > [card] [Community engagement](/intelligence/action-center) > > Build a queue of high-value posts to engage with and track them in the Action Center. > [card] [Sentiment research](/experiments/sentiment) > > Understand how users describe your category to inform sentiment experiments. # Introduction Source: https://docs.tryxlr8.ai/introduction Welcome to XLR8.ai — the AI Search Optimization platform for brands that want to win in the age of AI-powered search. XLR8.ai helps you track, analyze, and improve your brand's presence across AI search engines and large language models. As more users turn to ChatGPT, Perplexity, Gemini, Claude, and other AI tools to discover products and services, your visibility in those AI responses directly impacts your business. > [card] [Quick Start](/quickstart) > > Create your first experiment and start tracking your AI visibility in minutes. > [card] [Core Concepts](/concepts) > > Understand experiments, executions, models, and how XLR8.ai works. > [card] [Experiments](/experiments/overview) > > Run visibility and sentiment experiments across multiple LLMs simultaneously. > [card] [Analytics](/analytics/overview) > > Track how your AI visibility and sentiment change across executions. > [card] [Reporting](/reporting/overview) > > Review Search Console, Google Analytics, and first-touch survey data. > [card] [Content](/content/overview) > > Generate, optimize, and publish content designed for AI search. > [card] [Outreach](/outreach/overview) > > Turn visibility citations into listicle outreach campaigns. > [card] [Intelligence](/intelligence/action-center) > > Get AI-powered insights, action items, and social listening signals. > [card] [Integrations](/integrations/settings) > > Connect Webflow, WordPress, indexing, and outreach email. ## What is AEO? **Answer Engine Optimization (AEO)** is the practice of optimizing your brand's content and online presence so that AI-powered search engines and large language models include your brand in their responses. Traditional SEO focuses on ranking in Google's blue links. AEO focuses on being cited, mentioned, and recommended by AI assistants — the new discovery layer for millions of users. XLR8.ai gives you the tools to: * **Measure** how often and how favorably AI models mention your brand * **Understand** the sentiment and context around those mentions * **Optimize** your content so AI models recommend you more often * **Monitor** how changes over time affect your AI visibility * **Act** on data-driven recommendations across content, outreach, and indexing ## Supported AI Models XLR8.ai tracks your brand across leading AI search engines and assistants. The model picker in the app is loaded from the platform API; the frontend currently exposes these selectable models: | Model | | - | | Perplexity | | GPT Fast (GPT-5.1) | | GPT Thinking (GPT-5.1) | | GPT Fast (GPT-5.3 Chat Latest) | | GPT Thinking (GPT-5.4) | | Grok 4 | | Grok 4.20 | | Gemini | | Claude Opus 4.8 | | Claude Opus 4.7 | | Claude Sonnet 5 | | Claude Sonnet 4.6 | | Claude Sonnet 4.5 | | Google AI Mode | | Microsoft Copilot | > [!NOTE] > > The exact list shown in **New Experiment** comes from `/experiments/llm-models` and may update as new models are added. # Connect the MCP server Source: https://docs.tryxlr8.ai/mcp/connect Use XLR8.ai from Claude, ChatGPT, Cursor, Claude Code, VS Code, or Codex. Add the server URL, sign in with your XLR8.ai account, and start asking. The XLR8.ai MCP server lets your AI assistant work in your XLR8.ai account. Ask it to build a query set, start an experiment, break down which sources AI models cite, draft on-brand content, or find Reddit threads to join, all from the chat you already use. To connect, add the server URL to your assistant and sign in with the same account you use at [app.tryxlr8.ai](https://app.tryxlr8.ai). There is nothing to install. ```text theme={null} https://mcp.tryxlr8.ai/mcp ``` > [card] [Claude](#claude) > > Web, desktop, and Cowork > [card] [ChatGPT and Codex](#chatgpt-desktop-and-codex) > > ChatGPT desktop app, Codex CLI, and IDE extension > [card] [Cursor](#cursor) > > One-click install > [card] [Claude Code](#claude-code) > > One terminal command > [card] [VS Code](#vs-code) > > With GitHub Copilot > [card] [Other clients](#other-clients) > > Any client that supports remote MCP ## Set up your client ### Claude Add XLR8.ai as a custom connector. It works in Claude on the web, Claude Desktop, and Cowork. For more detail, see [Claude's custom connector guide](https://support.claude.com/en/articles/11175166-getting-started-with-custom-connectors-using-remote-mcp). > [tab] Free, Pro, and Max > > **1. Add a custom connector** > > Go to **Customize → Connectors**, select **+ Add**, then select **Add custom connector**. > > **2. Enter the server details** > > Name the connector `XLR8.ai`, paste `https://mcp.tryxlr8.ai/mcp`, and select **Continue**. > > **3. Sign in** > > Keep the authentication settings Claude detects, choose **Sign in now**, and select **Add**. Sign in with your XLR8.ai account. > > **4. Turn it on in a chat** > > In a conversation, select **+**, open **Connectors**, and turn on XLR8.ai. > > Free plans can add one custom connector. > [tab] Team and Enterprise > > An owner adds the connector once for the organization. Each member then connects their own XLR8.ai account. > > **1. Owner: add the connector** > > Go to **Organization settings → Connectors**, select **Add**, hover over **Custom**, and select **Web**. Name it `XLR8.ai`, paste `https://mcp.tryxlr8.ai/mcp`, keep the detected authentication settings, and select **Add**. > > **2. Member: connect your account** > > Go to **Customize → Connectors**, find XLR8.ai, and select **Connect**. Sign in with your XLR8.ai account. > > **3. Turn it on in a chat** > > In a conversation, select **+**, open **Connectors**, and turn on XLR8.ai. ### ChatGPT desktop and Codex The ChatGPT desktop app, the Codex CLI, and the Codex IDE extension share one MCP configuration. Add XLR8.ai in one of them and it is available in all three. For more detail, see the [Codex MCP documentation](https://developers.openai.com/codex/mcp/). > [tab] ChatGPT desktop app > > **1. Add the server** > > Open **Settings → MCP servers** and select **Add server**. Name it `xlr8`, choose **Streamable HTTP**, and paste `https://mcp.tryxlr8.ai/mcp`. > > **2. Restart and sign in** > > Save the server and select **Restart**. Then select **Authenticate** next to `xlr8` and sign in with your XLR8.ai account. > [tab] Codex CLI > > **1. Add the server** > > ```bash theme={null} > codex mcp add xlr8 --url https://mcp.tryxlr8.ai/mcp > ``` > > **2. Sign in** > > ```bash theme={null} > codex mcp login xlr8 > ``` > > Your browser opens so you can sign in with your XLR8.ai account. > [tab] Codex IDE extension > > **1. Add the server** > > Open the gear menu, select **MCP servers**, then select **Add server**. Name it `xlr8`, choose **Streamable HTTP**, and paste `https://mcp.tryxlr8.ai/mcp`. > > **2. Restart and sign in** > > Save the server and select **Restart extension**. Then select **Authenticate** and sign in with your XLR8.ai account. ### Cursor For more detail, see the [Cursor MCP documentation](https://cursor.com/docs/mcp). **1. Add the server** [Add XLR8.ai to Cursor](cursor://anysphere.cursor-deeplink/mcp/install?name=xlr8\&config=eyJ1cmwiOiJodHRwczovL21jcC50cnl4bHI4LmFpL21jcCJ9) and confirm the install when Cursor asks. To add it by hand, put this in `~/.cursor/mcp.json` to use it in every project, or in `.cursor/mcp.json` for one project: ```json theme={null} { "mcpServers": { "xlr8": { "url": "https://mcp.tryxlr8.ai/mcp" } } } ``` **2. Sign in** Open **Customize** in Cursor's sidebar, turn on XLR8.ai, and sign in with your XLR8.ai account when your browser opens. ### Claude Code For more detail, see the [Claude Code MCP documentation](https://docs.anthropic.com/en/docs/claude-code/mcp). **1. Add the server** ```bash theme={null} claude mcp add --transport http xlr8 https://mcp.tryxlr8.ai/mcp ``` To use XLR8.ai in every project, not only the current one, add `--scope user` after `xlr8`. **2. Sign in** Start Claude Code, run `/mcp`, select `xlr8`, and sign in with your XLR8.ai account in the browser. ### VS Code Use XLR8.ai with GitHub Copilot in VS Code. For more detail, see the [VS Code MCP documentation](https://code.visualstudio.com/docs/agent-customization/mcp-servers). **1. Add the server** Open the Command Palette (`Cmd+Shift+P` on Mac, `Ctrl+Shift+P` on Windows and Linux) and run **MCP: Add Server**. Choose **HTTP**, paste `https://mcp.tryxlr8.ai/mcp`, and enter `xlr8` as the server ID. **2. Choose where to save it** Select **Global**, then **Copilot Global**, to use XLR8.ai in every workspace. **3. Sign in** When VS Code asks, allow the sign-in and sign in with your XLR8.ai account. ### Other clients Any client that supports remote MCP servers can connect. Use these settings: * **URL:** `https://mcp.tryxlr8.ai/mcp` * **Transport:** Streamable HTTP (some clients call it HTTP) * **Authentication:** OAuth. Sign in with your XLR8.ai account when the client asks. If your client takes a JSON configuration: ```json theme={null} { "mcpServers": { "xlr8": { "url": "https://mcp.tryxlr8.ai/mcp" } } } ``` If your client can only run local servers, connect through [mcp-remote](https://www.npmjs.com/package/mcp-remote). It needs Node.js and opens your browser so you can sign in. ```json theme={null} { "mcpServers": { "xlr8": { "command": "npx", "args": ["-y", "mcp-remote", "https://mcp.tryxlr8.ai/mcp"] } } } ``` ## Check the connection Start a new chat and ask: Test the connection List my XLR8.ai organizations. The assistant should list each organization you belong to and your role in it. If you belong to more than one, tell it which one to use, for example "Use Acme for this chat." ## What you can ask Copy one of these prompts and replace the text in brackets. **Start a visibility experiment** Create a visibility query set for \[company] at \[domain] with about 20 queries. Show me the categories and queries first. After I approve them, save the query set and start an experiment with it. **See which sources AI models cite** For the experiment \[experiment name], group the cited sources by category and tell me which third-party sites appear most often. Suggest where we should work on getting mentioned. **Draft on-brand content** Write a blog post about \[topic] for \[organization]. Follow our brand guidelines and add internal links to our site. **Check a draft for AI detection** Scan each section of \[content title] with GPTZero. For any section that reads as AI-written, show me a humanized version and wait for my approval before you save it. **Find Reddit threads to join** Find recent Reddit threads where people ask about \[topic]. Pick the three most relevant and draft a helpful reply for each. > [!TIP] > > Some requests start work in your account, such as running an experiment, generating content, or sending an outreach email. Keep tool approval on in your assistant so you can review each action before it runs. ## Connect with an API key Use an API key when your client can't open a browser to sign in, for example an agent that runs on a server. Contact XLR8.ai to get an API key for your account. The client sends the key as a bearer token on every request: ```http theme={null} Authorization: Bearer ``` ```json Cursor theme={null} { "mcpServers": { "xlr8": { "url": "https://mcp.tryxlr8.ai/mcp", "headers": { "Authorization": "Bearer ${env:XLR8_API_KEY}" } } } } ``` ```bash Claude Code theme={null} claude mcp add --transport http xlr8 https://mcp.tryxlr8.ai/mcp \ --header "Authorization: Bearer $XLR8_API_KEY" ``` ```toml Codex theme={null} [mcp_servers.xlr8] url = "https://mcp.tryxlr8.ai/mcp" bearer_token_env_var = "XLR8_API_KEY" ``` ```json Other clients theme={null} { "mcpServers": { "xlr8": { "url": "https://mcp.tryxlr8.ai/mcp", "headers": { "Authorization": "Bearer " } } } } ``` The Cursor, Claude Code, and Codex examples read the key from an `XLR8_API_KEY` environment variable, so the key stays out of your config files. Set that variable before you start the client. If your client doesn't fill in environment variables, paste the key in place of the variable. In Claude, add the custom connector as described above, choose **No sign in**, and add a request header named `Authorization` with the value `Bearer `. > [!WARNING] > > On Claude Team and Enterprise plans, everyone who uses a connector set up with an API key acts as the key's owner. Use sign-in instead so each member connects their own account. If your key isn't tied to one organization, tell the assistant which organization to use. The [System Reporting API](/api-reference/overview) key is a different key and doesn't work with the MCP server. ## Troubleshooting > [accordion] Sign-in fails or never finishes > > Remove XLR8.ai from your client, add it again, and sign in with the same account you use at [app.tryxlr8.ai](https://app.tryxlr8.ai). > > If you connect through `mcp-remote`, also delete the `.mcp-auth` folder in your home directory to clear its saved sign-ins, then try again. > [accordion] The assistant says it isn't authorized > > Your saved sign-in expired or is no longer valid. Disconnect XLR8.ai in your client's connector or MCP settings and connect again. > > With an API key, check that the header contains the key itself and not the name of the environment variable. Copy the key again with no extra spaces. > [accordion] XLR8.ai doesn't show up in the chat > > Make sure the server is turned on. In Claude, select **+** in the conversation and check **Connectors**. In Cursor, check **Customize**. If you edited a config file, quit the client completely and open it again. > [accordion] My organization is missing > > The assistant only sees organizations your account belongs to. Ask an admin to [invite you](/account/authentication#inviting-team-members), or sign in with the account that already has access. > [accordion] My company blocks the connection > > Your IT team may restrict external MCP servers. Ask them to allow `mcp.tryxlr8.ai`. On Claude Team and Enterprise plans, only owners can add custom connectors. ## FAQ > [accordion] Why does the assistant ask to run search\_tools and call\_tool? > > XLR8.ai has a large set of tools. The assistant starts with a few, uses `search_tools` to find the one your request needs, and runs it with `call_tool`. Expand the approval request to see which XLR8.ai tool it will run and what it will send. > [accordion] What can the assistant access? > > Only organizations your XLR8.ai account belongs to. In those organizations, it can work with query sets, experiments, citation results, content, brand guidelines, Reddit listening, outreach, and videos. > [accordion] How do I disconnect? > > Remove XLR8.ai from your client's connector or MCP settings. In Claude, go to **Customize → Connectors** and select **Remove**. ## Related > [card] [Experiments](/experiments/overview) > > Track how AI models talk about your brand. > [card] [Query generation](/query-generation) > > Build the query sets your experiments run. > [card] [Content generation](/content/generation) > > Create content designed to improve your AI visibility. > [card] [Social listening](/intelligence/social-listening) > > Find Reddit conversations worth joining. # Outreach Source: https://docs.tryxlr8.ai/outreach/overview Run listicle outreach from visibility experiment citations — sync prospects, discover contacts, draft emails, and track replies. Outreach helps you turn visibility experiment citations into listicle outreach campaigns. You sync prospect articles from experiment results, discover contacts, draft emails with AI, send through connected Gmail, and track follow-ups and replies. Navigate to **Execution → Outreach** in the sidebar. > [!NOTE] > > Outreach is available for **visibility experiments** only. *** ## Connect Gmail first Before you can send email, connect a sender account: 1. Go to **Settings → Integrations → Outreach Email** 2. Click **Connect Gmail** 3. Optionally set a **Sender name** — used in generated emails and as the From name. Leave it blank to derive the name from the email address. You can also edit the name on a connected account and click **Save name**. 4. Optionally set a **Client tagline** for this organization. **Fill template** uses it in the draft; if you leave it empty, the template keeps `[Client tagline]` for you to replace. Click **Save tagline**. 5. Optionally set a **Default** connected account *** ## Hub tabs On the Outreach hub you can switch between: * **Experiments** — pick a visibility experiment to manage its outreach * **Outbox** — sent outreach emails across experiments * **Inbox** — replies across experiments **Inbox** and **Outbox** each have their own page. *** ## Inbox threads Click a reply in **Inbox** to open that conversation on its own page. You can reload the page and return to the same thread. On each row, set the thread to **Active**, **Inactive**, or **Done**. Use the status filter (**All**, **Active**, **Inactive**, **Done**) to show only matching conversations. Use **Mark as read** on an inbox row or the open thread to clear new replies. After a thread is read, **Mark as unread** puts it back so you can return to it later. *** ## Run outreach for an experiment **1. Open an experiment** From the **Experiments** tab, select a visibility experiment. **2. Sync prospects** Click **Sync** to discover listicle opportunities from experiment citations. Use **Add prospect** if you want to add one manually. **3. Find contacts** For each prospect, use **Find** to discover contacts with AI, or **Add** / **Manage** contacts manually. **4. Draft the email** Click **Write** or **Edit** in the Email column. Choose how to draft the email: **Generate from article**, **Write custom message**, or **Fill template**. Use the `[contact_name]` token in the body when you want personalization. When you generate from the article, choose **Generate as** to pick which connected mailbox the draft should use. The sender name in the draft comes from that mailbox. You can generate again after changing the mailbox. **Fill template** inserts the XLR8 listicle template immediately, with sender name, company, and article already filled in. It uses the **Client tagline** from Outreach Email settings, or `[Client tagline]` if you have not saved one. **5. Send** Click **Send**, choose the Gmail account under **Send from**, and set the follow-up schedule. Optionally turn on **CC recipients** to copy teammates (from your organization or another address) on the initial email and all automatic follow-ups. Follow-ups support up to 2 additional emails after 1–14 working days (weekends skipped). **6. Track status** Monitor prospect statuses: Pending, Sending, Sent, Follow-up sent, Reply received, Failed, or Inactive. *** ## CC after sending After outreach is scheduled or sent, you can still change **CC recipients** from prospect **Details** or the conversation thread. People you add are copied on the next follow-up or reply; a reply you send now includes them. When you reply, turn on **Include last message** if you want that reply to quote the previous email — useful when you CC someone mid-thread so they have context. *** ## Filters and summary chips On the per-experiment outreach table you can: * Search by domain or article * Sort by relevance, recently added, recently cited, most cited, or most competitors * Filter by min score, status, and assignee * Toggle **Show inactive** Summary chips show totals for Pending, Sent, Follow-up, Reply received, and Failed. *** ## Excluded domains In the experiment **Settings** modal, use the Outreach tab to exclude domains from prospect sync when you do not want those publishers contacted. *** ## Related > [card] [Visibility experiments](/experiments/visibility) > > Run the visibility experiments that feed outreach prospects. > [card] [Settings](/integrations/settings) > > Connect a mailbox and set the sender name and client tagline under Outreach Email. > [card] [Content generation](/content/generation) > > Create listicles and guides that improve your citation footprint. # Query Generation Source: https://docs.tryxlr8.ai/query-generation Build reusable query sets with company details, categories, and queries — then pick a set when you create an experiment. **Query Generation** is where you prepare the queries an experiment will run. Enter company details once, generate or import queries, and save a **query set**. When you create an experiment, you pick a saved set — you do not re-enter type, domain, aliases, or tracked competitors. From the sidebar under **Strategy**, click **Query Generation**. ## Query sets Each query set stores: * Experiment type and subtype * Company name, website, and aliases * Tracked competitors (inferred from domain analysis, or from competitor names on Competitor Alternatives and Competitor Comparison) * Categories and queries * Optional Keyword Planner keywords and search-volume targeting (language and locations), source citations, and comments The list shows company, type, category and query counts, and last update. Click anywhere on a row to open the set. Hover a row to download **CSV** or archive it (trash icon) — those actions stay on the list and do not open the set. Click **New query set** to start. If you already have sets, company fields can be prefilled from the most recent one. Use **Reuse company briefing** to copy type, company, domain, aliases, and Keyword Planner targeting from another set. ## Create a query set Creating a query set has three steps: **Company**, **Categories**, and **Queries**. **1. Company — analyze the domain** Fill in: * **Query set name** (optional — defaults from company and type) * **Experiment type** — Visibility Analysis or Sentiment Analysis * **Sub-type** — for visibility: Discovery, Informational, or Competitor Alternatives; for sentiment: Competitor Comparison or Brand Sentiment * **Company name** (required) * **Company website** (required) * **Categories & queries CSV** (optional) — upload a CSV with **Category** and **Query** columns to use your own queries instead of AI generation. Click **Download sample CSV** for the format, which includes an optional **Citations** column (max 5MB). Rows without citations are tagged from your briefing (domain analysis and additional context when those are filled), then enriched from uploaded documents, Keyword Planner data, and [Prospect Intelligence](/intelligence/prospect-intelligence) themes when company details are complete and analyzed calls exist. If you import before filling company details, add them and re-import to enrich tags. * **Additional context** (optional) — ICP, goals, known keywords, competitors, phrasing notes. When this organization has analyzed calls in [Prospect Intelligence](/intelligence/prospect-intelligence), a note under this field explains that query generation uses those structured themes automatically. * **Internal documents** (optional) — PDFs, Office files, text, or images; they are parsed and used when generating queries Click **Analyze domain**. XLR8.ai analyzes the website, prepares category suggestions, and infers **company aliases** and **tracked competitors**. Review and edit those on the next step — they are not entered on this step. **2. Categories — review and edit** For **Competitor Alternatives** and **Competitor Comparison**, this step is labeled **Competitors** instead of **Categories**. Review the **Domain description** and category (or competitor) chips. Add, rename, or remove items. If you have fewer than five, use **Generate categories** or **Fill to 5**. **Replace all from website** regenerates the full list from the domain. After domain analysis, review **Company aliases** (suggested from the site — add or remove chips) and **Tracked competitors**. For Competitor Alternatives and Competitor Comparison, tracked competitors match the competitor names in that list. For other subtypes, they are inferred from your company and domain and stay separate from query categories. Optionally use **Keyword Planner**: * Under **Search volume targeting**, pick a **Language** and one or more **Locations**. Both lists are searchable. Volume and competition data use these markets; the default is English and the United States. * **Fetch Keyword Planner ideas** using your categories plus extra seed keywords * **Add keywords with metrics** for exact keywords (comma or newline separated) * Remove any ideas you do not want to use Selected keywords and targeting are saved with the query set. Set **Queries per category** (1–20, default 5) to control how many queries are generated for each category. Click **Continue** to generate them. **3. Queries — generate and refine** A summary of **Company aliases** and **Tracked competitors** appears at the top — edit aliases on the Categories step. Review the query map. Use the pencil and trash icons on each row to edit or remove queries, or add a query to a category, then save the query set. Each query can show **citation** badges that explain where it came from, such as Domain, Document, Sales, Sales call, Context, Keyword, Model, Manual, Edited, or CSV. Click a badge to open **Why this query was generated**, with the supporting rationale. For **Document** sources, the panel lists **Source document** and links to download that file when the query set is saved. For **Sales call** sources, the panel lists the transcript (opens Prospect Intelligence), **Theme**, and **Type**. Queries without stored citation metadata show a **Manual** badge. Set **Queries per category** (1–20) and click **Regenerate queries** to run generation again with the current briefing, documents, Keyword Planner selection, and Prospect Intelligence themes. Queries you added or edited stay in the map; new suggestions are added around them. **Download CSV** exports the map (including citations when present). Click **Save query set**. After saving, comments become available on this step. ## Edit a saved query set Open a set from the list to change it. You land on **Queries**. Use **Back to categories** or **Back** to revisit categories and the company briefing. On the Company step, **Re-analyze domain** refreshes the description, category suggestions, aliases, and tracked competitors without losing the set. Keyword Planner keywords saved on the set refresh their monthly-search and competition metrics when you reopen it, using the language and locations stored on that set. ## Comments On a saved query set that has queries, a **Comments** sidebar appears on the Queries step. * Check a query to attach feedback to it, or leave a comment on the whole set. * Filter by **Open**, **Resolved**, or **All**. * **Post comment**, **Reply**, **Resolve**, or **Reopen** a thread. Notifications for query-set comments open that set. ## Use a query set in an experiment From **Strategy → Experiments**, click **New Experiment**. Step 1 is **Query set** — pick a saved set that already has queries. Experiment type, company, domain, aliases, and tracked competitors come from that set. Use **Manage query sets** or **Edit** if you need to change them first. Then continue to **Create** to choose models and launch. See [Creating a new experiment](/experiments/overview#creating-a-new-experiment). To change queries on an experiment that already exists, open **Settings → Queries** and click **Edit queries**. That opens the linked set here so you can add, edit, or remove queries. Changes apply to the next run; earlier results stay as they were. If the experiment is not linked to a set yet, **Edit queries** creates one from its current queries. > [card] [Experiments](/experiments/overview) > > Pick a saved query set and launch a visibility or sentiment experiment. > [card] [Visibility experiments](/experiments/visibility) > > Measure mention rate and citations across AI models. > [card] [Sentiment experiments](/experiments/sentiment) > > Analyze how AI models describe and compare your brand. > [card] [Core Concepts](/concepts) > > How query sets, category maps, and experiments fit together. # Quick Start Source: https://docs.tryxlr8.ai/quickstart Run your first AI visibility experiment and start tracking your brand across LLMs in minutes. This guide walks you through creating and running your first experiment on XLR8.ai. **1. Sign in and set up your organization** Go to [app.tryxlr8.ai](https://app.tryxlr8.ai) and sign in. If you're new, create an account and set up your organization. Accept the Terms of Service when prompted, then use the organization switcher in the header to select or create your organization. Open **Settings** and save your **Company Name**, **Company Domain**, and optional **Internal Team** names. XLR8.ai pre-fills company name and domain across experiments, content tools, and social listening. Internal Team names help [Prospect Intelligence](/intelligence/prospect-intelligence) keep prospect turns and skip your team's speech. **2. Create a query set** From the sidebar under **Strategy**, click **Query Generation**, then **New query set**. Enter your company name and website, choose **Visibility Analysis** or **Sentiment Analysis** and a subtype, then click **Analyze domain**. Review the suggested categories, aliases, and tracked competitors, optionally set **Queries per category**, click **Continue** to generate queries, edit them if needed, and **Save query set**. See [Query Generation](/query-generation) for CSV import, Keyword Planner, and comments. **3. Create your first experiment** From the sidebar under **Strategy**, click **Experiments**, then **New Experiment**. Creating an experiment has two steps: 1. **Query set** — pick the saved set (type, company, domain, aliases, and tracked competitors come from it) 2. **Create** — select AI models and optionally set a location Click **Create Experiment** to save the configuration and start the first run. **4. Wait for the execution to finish** On the experiment detail page, the execution status updates as it progresses. When it completes, open the results tabs to review the run. For visibility experiments, start with **Overview**, then dig into **Query Analytics** and **Source Analytics**. For sentiment experiments, use **Overview** and **Detailed Analysis**. **5. Explore insights and next actions** Use these follow-on tools with your first experiment: * **Insights** — ask natural language questions about the experiment data * **Analytics** — compare multiple executions over time once you have more than one run * **Action Center** — sync and track recommended fixes from sentiment results ## What's next? > [card] [Set up scheduling](/experiments/scheduling) > > Automate your experiments to run on a recurring cadence. > [card] [Generate optimized content](/content/generation) > > Create content designed to improve your AI visibility. > [card] [Review action items](/intelligence/action-center) > > See prioritized recommendations based on your experiment results. > [card] [Connect publishing](/integrations/webflow) > > Publish optimized content to Webflow or WordPress. # Reporting Source: https://docs.tryxlr8.ai/reporting/overview Connect Google Search Console and Analytics, upload monthly first-touch surveys, and review search, traffic, and attribution in one place. **Reporting** brings live Google Search Console and Google Analytics data together with monthly first-touch survey uploads. Use it to see search performance, site traffic, and how people say they found you. This is separate from [Analytics](/analytics/overview), which compares experiment executions over time. Navigate to **Reporting → Reporting** in the sidebar. *** ## Reporting vs Analytics | | Reporting | Analytics | | - | - | - | | Data | Search Console, Google Analytics, and survey CSVs | Experiment executions over time | | Best for | Search performance, site traffic, and first-touch sources | Visibility and sentiment trends across runs | | Location | **Reporting → Reporting** | **Reporting → Analytics** | *** ## Connect Google One Google connection covers both Search Console and Analytics properties. **1. Start the connection** Click **Connect Google** and complete Google authorization for the account that owns your properties. **2. Choose properties** Select a **Search Console** property and an **Analytics** property. The connector shows **Connected** once both are selected. **3. Refresh data** Charts load for the selected properties. Use **Refresh** to pull the latest data. If authorization is cancelled or expires, start **Connect Google** again. Connection errors appear at the top of the page and in place of charts, so a failed source does not look like zero traffic. Use **Disconnect** on the Google Data API card to remove the connector. **Show** / **Hide** next to Search Console or Analytics hides that source's charts without disconnecting. *** ## Filters After Google is connected, use the **Filters** bar and click **Apply filters**. **Reset** returns to the defaults. **Date range** * **Last X days** — Last 7, 28, or 90 days (default **Last 28 days**) * **Custom range** — pick a start and end date (the latest available day is yesterday) **Search Console** * **Device** — All devices, Desktop, Mobile, or Tablet * **Search type** — Web (default), Image, Video, News, Discover, or Google News * **Country** — optional three-letter code (for example, `usa`) **Analytics** * **Channel** — all channels, or a default channel group such as Organic Search or Direct *** ## Search Console When Search Console charts are visible, Reporting shows: * **Clicks**, **Impressions**, **CTR** (clicks / impressions), and **Avg position** (impression-weighted) * **Search clicks & impressions** — daily performance * **CTR & average position** — the position axis is inverted (lower is better) * **Top queries** and **Top pages** — ranked by clicks (lists start with a preview; use **Show all** to expand) * **Device split** — clicks vs impressions * **Top countries** — by clicks *** ## Google Analytics When Analytics charts are visible, Reporting shows: * **Sessions** and **Users** * **Site traffic over time** — sessions and users * **Traffic by channel** — default channel group * **Top landing pages** — by sessions, with users and engagement rate * **Source / medium** — sessions by source and medium * **Bounce rate over time** *** ## Unified page performance When both Search Console and Analytics charts are visible, **Unified page performance** joins Search Console pages and Analytics landing pages by path. Multiple Search Console URLs for the same path are grouped. If there are no exact path matches in the selected range, the table explains that no pages could be joined. *** ## Primary attribution **Primary attribution** is monthly survey data — typically “how did you hear about us” responses. It does not require the Google connection. **1. Open Manage data** On the Primary attribution card, click **Manage data**. If you have no uploads yet, you can also click **Upload CSV**. **2. Pick a month** Choose the calendar month the file belongs to. Uploading the same month again replaces that month's data. **3. Upload a CSV** Required columns are **Answer** (the source) and **Count**. A **Share** column is optional and ignored. Use **Download sample CSV** for the format, then drop a file or click to browse. Uploaded months appear in the dialog with the file name and response count. You can delete a month from that list. When data is present, Reporting shows: * **Total responses**, **Months uploaded**, and **Distinct sources** * **Responses by source** — source trends month over month * **Responses by month** — one dataset per calendar month Use **Show** / **Hide** on the Primary attribution card to hide these charts without deleting uploads. *** ## Related > [card] [Analytics](/analytics/overview) > > Compare experiment executions over time — a different view from this page. > [card] [Indexing](/indexing/overview) > > Submit URLs to Google Search Console and Bing Webmaster. > [card] [Settings](/integrations/settings) > > Company profile and other third-party integrations. > [card] [Experiments](/experiments/overview) > > Track AI visibility and sentiment across models.