# AgentData

> Free company intelligence for AI agents and growth teams: tech stack, email addresses with verification status, web signals and decision-makers for any company, via web, REST API or MCP server. Free forever; company, tech and signal endpoints need no key.

AgentData has indexed 133,126,016+ company domains and profiled 24M of them (name, tech stack, email pattern, signals; any other domain is profiled on request, usually within a few minutes), with 16,309,721+ email addresses found (every address carries a verification_status), 4169+ technologies detected and 5,612,560+ people profiles, refreshed continuously by Ady, its crawling agent. The live data page (https://agentdata.run/data) shows coverage, verification and freshness. Limits (free account): 1,000 contact reveals per day, 5,000 requests per 5 hours, 5 requests per second; without a key, company, tech and signal endpoints allow 300 requests per 5 hours and 200 distinct companies a day per IP. Contact reveals reset at midnight UTC; a company you have already revealed stays open for 30 days without using another. Pro ($99/year): 10,000 reveals per day. Built in the UK by Nick and Duda Timms, whose commercial products are Drag (dragapp.com, shared inboxes for Gmail, 240,000+ users) and HeyHelp (heyhelp.ai, AI email assistant for Gmail), which is why AgentData is free. Operated by DragApp.com Ltd, registered in England and Wales (company number 10835737).

## Start here

- [API documentation](https://agentdata.run/docs): endpoints, authentication, rate limits and examples for the REST API
- [Sign up](https://agentdata.run/signup): free account with an API key issued on first login
- [MCP server guide](https://agentdata.run/docs/mcp): connect AgentData to ChatGPT, Claude, Claude Code, Cursor, VS Code or any MCP client; tools, limits and privacy
- [MCP server](https://mcp.agentdata.run/mcp): hosted Streamable HTTP endpoint for ChatGPT, Claude, Cursor and other MCP clients; no key needed for company, technology and signal tools (details below)

## MCP server

Hosted at https://mcp.agentdata.run/mcp (Streamable HTTP). Read-only tools. Without a key the tools give what agentdata.run shows to a visitor who is not signed in: lookup_company (the public company page: facts, tech stack, email format, general inboxes, people's titles and LinkedIn URLs, activity, changes, similar companies), search_companies, find_people (the first 20 matches, as on the public people directory), get_technologies, get_signals and check_usage. Personal email addresses (lookup_company with a key, get_person) and full people results need a free API key (no card) sent as `Authorization: Bearer YOUR_API_KEY`. Every response ends with a usage line so an agent can pace itself.

- [Add to Claude Code without a key](https://agentdata.run/docs): `claude mcp add --transport http agentdata https://mcp.agentdata.run/mcp`
- [Add to Claude Code with a free key](https://agentdata.run/settings): `claude mcp add --transport http agentdata https://mcp.agentdata.run/mcp --header "Authorization: Bearer YOUR_API_KEY"`
- [ChatGPT plugin](https://chatgpt.com/plugins/plugin_asdk_app_6aba5e34deb88191bbb7b8b09d5c5f78): in ChatGPT's plugin directory as AgentData (Plugins, search AgentData, Install plugin); no sign-in; the keyless tools (company profiles, search, technologies, signals, people's titles)
- [claude.ai custom connector](https://agentdata.run/docs/mcp): add the URL with no auth; the keyless tools work there (company profiles, search, technologies, signals, people's titles)
- [MCP server source (GitHub)](https://github.com/nick-timms/agentdata-mcp-server): tools, limits, example prompts and a local stdio option (npm package agentdata-mcp-server)

## Browse the data

- [Companies](https://agentdata.run/companies): search and filter companies by sector, size, tech and signals
- [Company directory by sector](https://agentdata.run/companies/directory): every sector and vertical with counts
- [Technologies](https://agentdata.run/tech): every detected tool, with the companies using it
- [Technology directory](https://agentdata.run/tech/directory): tools grouped by category
- [People](https://agentdata.run/people): decision-makers and team members with roles
- [Local businesses](https://agentdata.run/local-businesses): local business listings by category and city
- [Local business directory](https://agentdata.run/local-businesses/directory): all categories with counts

## Page patterns

- [Company page](https://agentdata.run/company/stripe.com): /company/{domain}, one page per company with tech stack, emails, people and signals
- [Technology page](https://agentdata.run/tech/zendesk): /tech/{slug}, companies using a tool, paginated at /tech/{slug}/page/{n}
- [Technology alternatives](https://agentdata.run/tech/zendesk/alternatives): /tech/{slug}/alternatives, tools that replace it
- [Local business category](https://agentdata.run/local-businesses/jiu-jitsu/calgary): /local-businesses/{category}/{city}

## Optional

- [Live data](https://agentdata.run/data): what we have, how complete and how fresh it is, and how it was collected
- [Privacy policy](https://agentdata.run/legal/privacy)
- [Data removal](https://agentdata.run/legal/data-removal): how a company or person asks to be removed