Onplana

What is Onplana?

Onplana provides open-source TypeScript Model Context Protocol building blocks extracted from Onplana’s production MCP deployment. It includes a server template (Streamable HTTP transport, Bearer auth, prompt‑injection containment, pluggable dispatcher) and a typed TypeScript client SDK for calling the public Onplana MCP endpoint at https://api.onplana.com/api/mcp/v1.

How to use Onplana?

Install the server template or client SDK via npm install github:Onplana/onplana-mcp-server. For a custom server, wire an Express app with createMcpPostHandler, requireBearerAuth, and your own Dispatcher. For driving Onplana, instantiate OnplanaMcpClient with an Onplana PAT and call methods like listProjects or searchOrgKnowledge.

Key features of Onplana

• Prompt‑injection containment via wrapUserContent + escape utilities.
• Streamable HTTP transport with stateless mode.
• Scoped Bearer authentication.
• Pluggable Dispatcher interface for business logic.
• Compatible with Claude Desktop, Cursor, ChatGPT, Gemini CLI, GitHub Copilot, MCP Inspector.
• Client SDK for Onplana’s hosted MCP endpoint.

Use cases of Onplana

• Build a secure, production‑grade MCP server from the template.
• Drive Onplana project‑management tools from AI agents.
• Search across org knowledge with hybrid semantic + lexical search.
• Use as a reference implementation for MCP security patterns.

FAQ from Onplana

What runtime does Onplana require?

Node.js ≥ 20 for the server template, ≥ 18 for the client (uses ambient fetch). Requires @modelcontextprotocol/sdk@^1.29.0 and express@^4.18.0 or ^5.0.0.

How do I install Onplana in Gemini CLI?

Set ONPLANA_PAT and run gemini extensions install https://github.com/Onplana/onplana-mcp-server. Restart the Gemini CLI or reload your editor.

What authentication and transport does the server use?

The server template implements Streamable HTTP transport and scoped Bearer authentication. Clients pass a token validated by a custom auth function.

Where does user data reside?

Data stays in Onplana’s platform. The client SDK calls the public Onplana MCP endpoint; the server template runs your own dispatcher and can store data anywhere you choose.

Are there any known limits?

The README production checklist recommends per‑token rate limiting (60–120 req/min), tenant cost caps, and idempotency keys, but the template itself imposes no hard limits.