Docana MCP

Docana is a hosted MCP server. Point your AI tools at it and they can work your workspace directly: search your knowledge, upload documents, manage collections, review conversations, and build agents — all in plain language. Nothing to install, nothing to run. Works with Claude Code, Claude Cowork, Claude Desktop, Cursor, and any client that speaks MCP.

:::info Looking for the other direction? This page is about using Docana from your AI tools. To give your Docana agents tools from external MCP servers, see MCP Servers. :::

Connect

The server lives at https://platform.docana.com/mcp and authenticates with a Docana API key in the Authorization header, using the ApiKey scheme (not Bearer). Use a Full access key — connected tools can do exactly what the key can do, nothing more.

Claude Code

claude mcp add --transport http docana https://platform.docana.com/mcp \
  --header "Authorization: ApiKey YOUR_API_KEY"

Cursor

Add to .cursor/mcp.json (or via Settings → MCP):

{
  "mcpServers": {
    "docana": {
      "url": "https://platform.docana.com/mcp",
      "headers": { "Authorization": "ApiKey YOUR_API_KEY" }
    }
  }
}

Claude Desktop

Claude Desktop launches local servers, so bridge with mcp-remote in claude_desktop_config.json:

{
  "mcpServers": {
    "docana": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote", "https://platform.docana.com/mcp",
        "--header", "Authorization: ApiKey YOUR_API_KEY"
      ]
    }
  }
}

Other clients

Any MCP client that supports remote servers over Streamable HTTP works:

• URL: https://platform.docana.com/mcp
• Header: Authorization: ApiKey YOUR_API_KEY

Things to ask for

Once connected, just talk — your AI tool picks the right Docana capabilities:

• "Search my knowledge for our refund policy and summarize it."
• "Create a collection called Contracts and upload this PDF into it."
• "Upload https://example.com/handbook.pdf to the Onboarding collection."
• "List my applications and show yesterday's conversation threads for the support app."
• "Build an agent that triages support tickets by urgency and hands the urgent ones to a human."
• "Validate this agent spec and tell me what's wrong with it."
• "Rename document 4231 to signed-contract and archive the old draft."

What your AI tool can do

Capability	What's in it
Tools	The full platform — knowledge, documents, collections, conversations, agents, applications, search. Reads are marked read-only and deletes destructive, so well-behaved clients ask before the risky ones.
Resources	Your content as attachable context: collections (with their documents), document files, conversation transcripts, and the agent-building guide.
Prompts	build-agent — the same expertise Docana's own agent builder works with, so your AI tool designs agents like Docana does. upload-document — teaches the right upload path for where a file lives.

Uploading files

Uploads just work — the model picks the right path for where the file lives:

• A URL — Docana downloads it (https:// links, up to 25MB).
• Text it already has — sent inline.
• A local file — sent directly as raw bytes through a one-time upload link (up to 30MB), so binaries like PDFs upload cleanly.

Building agents

Ask your client to load the build-agent prompt — or just ask it to build an agent. It designs the agent, validates the result until it's clean, imports it into your application, and verifies it — the same pull-edit-push rhythm the CLI offers with files.

Troubleshooting

Symptom	Fix
401 on every call	The header scheme is ApiKey, not Bearer: Authorization: ApiKey YOUR_KEY.
Tools exist but calls return 403	The key lacks access — Widget-only and Routine-only keys are locked to their endpoints. Use a Full access key.
A recently added capability doesn't show up	MCP clients cache the tool list per connection. Reconnect the server (or restart the session).

Next Steps

• Generating API Keys - The key the MCP authenticates with
• API Overview - Integrate from your own backend
• CLI - Work your agents from the terminal