MCP docs › FAQ
Frequently asked questions
What is the ConvertFleet MCP server?
A remote Model Context Protocol server at https://convertfleet.online/mcp that lets any MCP-compatible AI agent run ConvertFleet's web-data tools — Google Maps lead scraping, ad-library intelligence, web search and page reading, social and e-commerce extraction, YouTube data and transcripts, email finding and verification, and media download links.
Which AI agents can connect?
Anything that speaks MCP over Streamable HTTP: Hermes, Claude (Code and Desktop), Cursor, OpenClaw, n8n, Zapier, Make, or your own client via the Python/TypeScript SDKs. A legacy SSE transport is available for older integrations.
How do agents know which tool to use?
The server sends usage instructions on connect that most clients place in the model's system prompt, every tool description states when to use it and what it is not for, and the find_tool helper maps a plain-English job to the right tool with example arguments. None of the helpers consume quota.
Why did a tool return status "running" instead of results?
Scrapes run as jobs. If one exceeds the inline wait you get a job_id back — poll get_job_result with that id until the status is done. It is a successful response, not an error.
Does an MCP call cost the same as using the dashboard?
Yes. MCP, REST API and dashboard runs all meter against the same monthly plan allowance, and usage is tracked per source so you can see what your agents consumed.
Can I limit what an agent can do?
Yes. Mint a scoped cfm- token and tick only the tools that agent needs. Out-of-scope tools are hidden from tools/list and rejected server-side, and any token can be revoked on its own.
Is an empty result a failure?
Usually not — it means nothing matched, which is a real answer. Runs that were actually blocked return an error with the reason instead.
Where is the machine-readable catalog?
https://convertfleet.online/mcp/manifest.json for the tool catalog with JSON Schemas, https://convertfleet.online/mcp/openapi.json for the endpoint, https://convertfleet.online/llms.txt and https://convertfleet.online/llms-full.txt for LLM ingestion, and https://convertfleet.online/mcp/docs.md for this documentation as Markdown.