For AI agents: a documentation index is available at /llms.txt. A markdown version of this page is available at /guide/mcp/introduction.md.
MCP / AI Connect
Connect Claude, Cursor, GitHub Copilot or the Codex CLI to Orangescrum and let your assistant read and change projects, tasks, timelogs and users.
The Developer API doubles as a Model Context Protocol server. Point any MCP-compatible client at it and your assistant gains a set of Orangescrum tools — it can list your projects, open a task, log time or file an update on your behalf, using the same partner API key you already use for REST.
#Endpoint
POST https://v4-api.orangescrum.com/mcp/partner
The server speaks JSON-RPC 2.0. Authenticate with your partner key in the
X-API-KEY header. Every active partner key works as an MCP token — there is no
separate provisioning step.
Two protocols, one key
This is the same credential as the REST API, and it draws on the same rate limits: 120 requests per minute, 5,000 per day. Give a busy assistant its own key so it cannot starve your integrations.
#What your assistant gets
CRUD plus search across projects, tasks, timelogs and users.
💬2 promptsdaily_standup and overdue_report, surfaced as slash commands.
#Supported clients
| Client | Transport | Config file |
|---|---|---|
| Claude Desktop | STDIO bridge via mcp-remote | claude_desktop_config.json |
| Cursor | Remote HTTP | ~/.cursor/mcp.json |
| VS Code + GitHub Copilot | Remote HTTP | .vscode/mcp.json |
| OpenAI Codex CLI | Remote HTTP | ~/.codex/config.toml |
Full snippets for each are on the Client setup page.
OAuth clients are not supported yet
Browser-based MCP clients that require an OAuth handshake — such as custom
connectors on claude.ai — cannot connect. The server authenticates with
X-API-KEY only, which covers every desktop and CLI client above.
#Is this safe to point at production?
Worth deciding deliberately before you connect it:
- The tools include writes. An assistant can create and update real tasks, projects and timelogs — not just read them.
- The key's scopes are the boundary. A read-only key gives you a research assistant with no ability to change anything; that is the right starting point.
- Every call lands in the audit trail with its status code and response time, so you can see exactly what the assistant did.
Tip
Start with a read-only key against a non-critical workspace, watch the audit log for a few days, then widen scope once you trust the workflows.