MCP & AI agents

Orkestra exposes ~290 MCP tools covering 99% of the backend. Any compatible MCP client (Claude Desktop, OpenClaw, HermesAgent, etc.) can operate Orkestra with natural language.

Supported transports

The Orkestra MCP server supports two transports:

  • STDIO (local) — the client starts the server as a subprocess. Useful for Claude Desktop, where the MCP runs alongside the client. Auth via MCP_TOKEN env var.
  • Streamable HTTP (remote) — the client connects to https://mcp.orkestra.team. Useful for agents running on servers (OpenClaw, HermesAgent, N8N). Auth via Authorization: Bearer ork_... header.

Generate an MCP token

Go to Settings → MCP Tokens → Create token. Give it a descriptive name (e.g. "Claude Desktop Juan's laptop") so you can revoke it later.

The token value is shown only once. Copy it and store it in a password manager. If you lose it, create a new one (and revoke the old).

Available tools

The ~290 tools are organized in 35 modules. The main ones:

  • Organizations (5 tools) — list, get, members, create, manage
  • Projects (9 tools) — full CRUD, stats, archive/restore
  • Tasks (10 tools) — CRUD, move state, assign, comment, subtasks
  • Wiki (13 tools) — CRUD + versions + comments
  • Workflows (5 tools) — states, transitions, reorder
  • Analytics (2 tools) — project metrics, org dashboard
  • Automation (5 tools) — automation rules
  • Audit (5 tools) — history, activity, alerts
  • Bulk (4 tools) — bulk operations (move, update, assign, delete)
  • Processes (5 tools) — process engine with approvals
  • Stock (15 tools) — physical inventory + budget
  • Products (10 tools) — business lines with KPIs
  • ...and 23 more modules (chat, notifications, webhooks, custom fields, time tracking, etc.)

Automatic preview / confirm

All write operations go through a preview/confirm flow. When an agent requests to create, update or delete something, the response is a "pending action" with a confirmToken. Only when the agent calls confirm_action with that token is the operation executed.

This prevents an agent from executing destructive changes by accident or hallucination. Compatible clients show a preview to the user before confirming.

Connecting clients

For step-by-step guides with copy-paste configuration, go to the full AI Agents page. It covers Claude Desktop, OpenClaw and HermesAgent with examples.

Limits and rate limiting

The MCP server respects the same rate limits as the REST API: 1000 requests per minute per token on normal operations, 100/minute on sensitive write operations (bulk, custom roles, webhooks). If you see 429, wait and retry with exponential backoff.

What's NOT in MCP (by design)

Some operations are intentionally excluded from MCP:

  • Billing — plan changes require UI with human confirmation.
  • MCP token management — avoid circular loops (a token can't create tokens).
  • Calendar sync OAuth — requires user redirection.
  • Export/Import with files — the MCP transport doesn't handle large binaries well.

Resources