Use first-party evidence to see whether AI cites your site, fetches your pages, and refers visitors back to you.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install seomatic-ai-citations
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install seomatic-ai-citations using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is SEOmatic AI Citations?
SEOmatic AI Citations helps you measure how AI platforms use your website using stored first-party evidence: citations observed in real, logged-in AI answers, server-side scan results, AI crawler fetches, and visits referred from AI assistants. It is an Openclaw Skills integration for investigating AI visibility without mistaking a clean, logged-out scan for what a real user saw.
The skill also checks canonical business facts and internal-link health to help explain inaccurate AI answers. Reads are free, scoped to the connected SEOmatic workspace, and do not trigger scans or spend credits. SEOmatic works from real Google Search Console data, and proposed site changes require user approval before going live.
SEOmatic AI Citations Use Cases
- Find out whether ChatGPT or another AI assistant actually cites the site.
- Measure which pages AI crawlers fetch and which pages receive AI-referred visits.
- Investigate pages that AI reads but does not cite or link to.
- Compare observed citations across AI engines, including citation rate and best position.
- Check published business facts when an AI answer describes the business incorrectly.
- Use server-side scan results as a clearly labeled fallback when no observed citation data is available.
How SEOmatic AI Citations Works
- Query
ai_citationswith theobserved_citationsaction to retrieve real answers captured by the user's browser extension. The default reporting window is 28 days. - Lead with observed data when available. Review citation rate, answers observed, citations, and best citation position for each engine. If there is no observed data, explain that the extension may not be installed or logging may be off; do not interpret an empty result as proof that AI never cites the site.
- If observed data is unavailable, use
ai_citationswithscan_resultsfor comparable server-side results from clean, logged-out prompts. Clearly identify these as scanned rather than observed results. - Query
traffic_analyticswith theai_pagesaction to see per-page AI crawler fetches (fetchedByAi) and AI-referred visits (aiVisits). Interpret the metrics together to distinguish pages AI reads and sends visitors to from pages it reads without linking. - When investigating incorrect business information, query
ai_citationswithbrand_facts. Compare the AI claim with published canonical facts and check internal-link health; treat missing authoritative facts as a possible reason for inaccurate answers. - Close with recommendations supported by the evidence, such as improving a readable but unquotable page or making missing business facts available. Never claim a citation that the data does not show.
SEOmatic AI Citations Setup
- Connect the SEOmatic MCP server to the AI agent. The server uses streamable HTTP at
https://app.seomatic.ai/api/mcpand OAuth authentication. - Alternatively, create a Bearer API key beginning with
smk_in the SEOmatic dashboard's AI agents settings:https://app.seomatic.ai/dashboard/settings?tab=ai-agents. - Make sure the SEOmatic MCP server is available to the agent. The skill requires the
seomaticMCP server; it does not require environment variables or local binaries. - To collect observed citations, install and enable the SEOmatic browser extension so it can capture real, logged-in AI answers. If observed data is not available, the skill can use scanned results instead.
All reads in this skill are free on the SEOmatic free tier. They use stored first-party data, are already scoped to the connected workspace, and do not initiate scans or consume credits. Do not ask the user for their domain.
SEOmatic AI Citations Data Schema & Taxonomy
The skill reads stored data through two tools:
| Tool | Action | Data returned | Interpretation |
|---|---|---|---|
ai_citations |
observed_citations |
Per-engine counts of observed answers and citations, citation rate, and best citation position; default window is 28 days | Evidence captured from real, logged-in AI answers by the user's browser extension |
ai_citations |
scan_results |
Server-side results for clean, logged-out prompts | De-personalized, comparable scan data; label it as scanned, not observed |
ai_citations |
brand_facts |
Canonical published business facts, official profiles, and an internal-link health reading | Use to compare AI claims with the site's declared facts |
traffic_analytics |
ai_pages |
Per-page fetchedByAi and aiVisits metrics |
Shows AI crawler fetches and visits referred from AI assistants |
No generated files or local data store are specified. The skill reads stored first-party data from the connected SEOmatic workspace. High fetches with high visits suggest AI uses a page and sends people to it; high fetches with few visits suggest the page may be readable but not worth citing; near-zero fetches may indicate an indexability issue or a page that does not answer a relevant question.
SEOmatic AI Citations Advanced Features
- Separates real-user observed citations from clean, logged-out server-side scans to prevent misleading comparisons.
- Reports citation rate, answer volume, and best citation rank by AI engine; a best position near 1 indicates a leading citation when the site is cited.
- Connects crawler fetches with AI-referred visits, exposing traffic that ordinary analytics may classify as Direct.
- Audits canonical business facts and internal-link health to help diagnose inaccurate AI answers.
- Uses workspace-scoped reads and requires no domain input, local binaries, or environment variables.
- Supports OAuth or a Bearer
smk_API key through the SEOmatic MCP server. - Provides read-only evidence gathering; site changes are not made live unless the user approves them, and SEOmatic changes can be rolled back.
SKILL.md
Loading
Related Openclaw Skills
SEOmatic Weekly SEO Brief
SEOmatic turns real Search Console data into a concise weekly SEO update that explains what changed and recommends one data-backed action.
SEOMatic AI Visibility Review
Review how AI answer engines represent your brand, refine the prompts you track, and report actionable visibility gaps using SEOMatic data.
SEOmatic SEO Audit
SEOmatic SEO Audit analyzes real Search Console, indexing, technical crawl, and AI-search data to produce a prioritized SEO fix list.
SeaGames CLI
SeaGames CLI lets Openclaw Skills generate, process, recover, and deliver creative assets across image, video, audio, text, and 3D workflows.
SEOmatic Approval Workflow
Review SEOmatic proposals against real Search Console evidence, record your decisions, monitor approved changes, and roll back live edits when needed.
SEOmatic Local SEO Review
Review a local business’s search visibility, Google Business Profile, and website signals, then stage data-backed fixes for the user’s approval.
Featured*
