Give your AI agents SEO and AI visibility.
Connect an agent to Trackee through MCP or a coding skill and let it answer questions about your rankings, AI mentions, citations, and competitors in plain language.
POST https://api.trackee.dev/v1/visibility
{ "brand": "Salesforce", "prompt": "best crm software", "engines": ["chatgpt"] }Two ways to connect
Use a skill inside your coding agent, or the MCP server for any compatible client. Both expose the same Trackee tools.
Agent skill
Add the Trackee skill to your coding agent so it knows the API and can call it directly.
npx skills add trackee-API/skills --allMCP server
Point any MCP client at mcp.trackee.dev/mcp and authenticate with your access key.
{
"mcpServers": {
"trackee": {
"url": "https://mcp.trackee.dev/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY"
}
}
}
}Built for agents, not one-off scripts
Small, predictable tools with clear costs and clean JSON, so an agent can plan work and report back.
Answer, do not just search
Your agent asks a real question, how does our brand show up in ChatGPT this week, and gets structured data back, not ten blue links to read.
One key, every check
Rankings, AI visibility, mentions, keywords, and competitors are all behind a single access key. No per-source setup.
Credits, not surprises
Every tool reports what it cost and what is left, so an agent can budget its own work and never runs blind.
Reads are free
Checking usage and managing brands never costs credits, so an agent can plan and organize before it spends anything.
The tools your agent gets
A small surface that mirrors the Trackee API.
Visibility and search
- get_visibility — does ChatGPT, Claude, Gemini, or Perplexity mention a brand
- get_rank — Google position and AI Overview presence for a keyword
- get_mentions — web pages that mention a brand for a query
- get_competitors — the domains competing for a set of keywords
- get_keyword_volume — search volume, CPC, and competition
Brands and account
- create_brand, list_brands, get_brand, update_brand, delete_brand
- set_brand_prompts, set_brand_competitors
- get_usage — credit balance and history (free)
- api_health — check the API is up (free)
One agent, every answer engine
Your agent does not need to learn each model. Trackee checks them for you.
- ChatGPT
- Claude
- Gemini
- Grok
- Perplexity
- Google
SDKs and tools
One API. Every way to build.
Build with the REST API, Node.js or Python SDKs, AI agents, or the dashboard. All use the same Trackee API.
REST API
One predictable contract for visibility checks, normalized results, and organization-scoped access.
AI agents
Let Claude, Codex, Cursor, or your own agent query Trackee through MCP and agent-ready tools.
Dashboard
Use the same underlying data in a visual workspace for teams that prefer to inspect and compare results.
const response = await fetch(
"https://api.trackee.dev/v1/track",
{
method: "POST",
headers: {
"x-access-key": process.env.TRACKEE_ACCESS_KEY,
"Content-Type": "application/json",
},
body: JSON.stringify({
brand: "Acme",
prompt: "best project management software",
engines: ["chatgpt", "claude", "google"],
}),
},
);
const visibility = await response.json();Frequently asked questions
Do I need to run anything locally?
No. The MCP server is hosted at mcp.trackee.dev/mcp. Add the URL to your AI client and you are connected. Nothing to install.
Which clients work?
Any MCP-compatible client, including Claude.ai, Claude Desktop, Claude Code, Cursor, and other Streamable-HTTP clients. See the MCP guide for exact config.
How does billing work through an agent?
The MCP server forwards every call to the same API, so credits, rate limits, and access-key permissions behave exactly like REST. Reads such as usage and brand management are free.
Can the agent manage brands?
Yes. It can create brands, set the prompts and competitors to track, and run visibility, rank, and mentions checks against them.
Full setup is in the MCP guide.