Agentic Platform
Documentation API reference Tools Skills

Connect

Set up MCP

The agents are exposed as MCP tools over streamable HTTP. Any MCP client — Cursor, Claude, the MCP Inspector — can list them, read their schemas and call them on your behalf.

Quick install

On Claude, ChatGPT, Perplexity, Gemini, or your preferred answer engine, ask: 

Please install the Botify MCP from https://mcp.botify.com

Endpoint

Transport is streamable-http. Point your client at this URL:

MCP https://mcp.botify.com

Authentication

OAuth 2.1, handled entirely by the client: it reads the endpoint's protected-resource metadata (RFC 9728), registers itself with Botify (RFC 7591) and runs an authorization-code flow with PKCE. You sign in and approve once in the browser; the client refreshes its own access token from then on.

Cursor

Add the server to ~/.cursor/mcp.json (or .cursor/mcp.json in a project), then reload the window and approve the connection in the browser window Cursor opens.

~/.cursor/mcp.json
{
  "mcpServers": {
    "botify-agents": {
      "url": "https://mcp.botify.com"
    }
  }
}

Claude

In Claude Desktop and on claude.ai, go to Settings → Connectors → Add custom connector and paste the endpoint.

In Claude Code, add the server from the terminal, then run /mcp in the session and pick it to sign in. 

claude mcp add --transport http botify-agents https://mcp.botify.com

Try it with the MCP Inspector

The official inspector is the quickest way to check the connection and browse the tools.

npx @modelcontextprotocol/inspector

In the inspector, choose the Streamable HTTP transport, paste https://mcp.botify.com and connect — it runs the same browser sign-in.

Tools available over MCP

52 tools are published to MCP. Each one takes an organization, a project, an item and an optional config. The list_projects tool is MCP-only and helps the model resolve the first two.

See them in the catalog →