DiffDeck API
Automate DiffDeck over a REST API, or connect it to any MCP client. Every action below is available through both — authenticated with a single API key.
REST API
Invoke any action with a POST request. Pass your key as a bearer token.
curl https://diffdeck.ai/api/rest/actions \
-H "Authorization: Bearer kav_..."POST https://diffdeck.ai/api/rest/actions/{action}MCP server
Point any Model Context Protocol client at the endpoint below. Each permitted action becomes a tool.
https://diffdeck.ai/api/mcpAuthorization: Bearer kav_...Let Claude work in DiffDeck for you
Add DiffDeck to Claude as a connector, and you can ask Claude to look things up or make changes for you in plain conversation — on claude.ai, the Claude app, or Claude Code. It signs in as you, so it only ever sees what you're allowed to see.
https://diffdeck.ai/api/mcpClaude (web, desktop, mobile)
Open Claude and go to Settings (or Customize Claude) → Connectors.
Choose 'Add custom connector' and paste in the address above.
Sign in with your DiffDeck account, pick which organization to connect, and approve — you can allow read-only access, or allow changes too.
Claude Code
Run this once in a terminal:
claude mcp add --transport http diffdeck https://diffdeck.ai/api/mcpThen run /mcp inside Claude Code and sign in when prompted.
Other MCP clients
Most MCP clients can sign in the same way (they'll open a browser tab to ask you to log in and approve). If yours can't, use an API key instead, exactly like the REST API above:
https://diffdeck.ai/api/mcpAuthorization: Bearer kav_...What Claude can (and can't) do
Claude can only do what your own role in DiffDeck allows — nothing more.
Choose read-only access if you only want Claude to look things up, never change them.
Every change Claude makes is recorded in the audit log, just like a change you made yourself.
You can disconnect it at any time from Settings → Connected apps.