Skip to main content
The Model Context Protocol (MCP) lets AI tools search and read this documentation directly while they answer you. Instead of relying on training data or a generic web search, an MCP-connected tool queries the current Zeam docs and grounds its answer in them.

​
What it is

This site is published with Mintlify, which hosts a search MCP server for the documentation. Any MCP-capable client (Claude, Claude Code, Cursor, VS Code, and others) can connect to it. The server is read-only: it searches and returns pages, and it never calls the Zeam API or changes anything. It exposes tools to:
  • search across all indexed pages, including this guide set and the API reference;
  • read full pages as Markdown;
  • report a documentation problem back to the Zeam docs team.
Because it searches the live indexed site, answers stay aligned with what is published here.
The hosted MCP server is the /mcp path of the docs site: https://docs.zeam.money/mcp.

​
Connect a client

Use the docs MCP URL https://docs.zeam.money/mcp (HTTP transport). The public docs server needs no authentication.
For Claude, add it under Settings → Connectors → Add custom connector with the same URL. You can also connect from any page of this site using the page’s contextual menu.

​
What you can ask

Once connected, ask your tool questions that it answers from these docs, for example:
  • “How do I authenticate with the Zeam API?” to find the token flow.
  • “Which endpoint lists my wallets?” to locate GET /v1/wallets.
  • “Generate a curl request for a P2P transfer.” to assemble a POST /v1/payments/p2p call.
  • “What does a failed transaction status mean and how do I see the reason?”
  • “Why am I getting a 412 from the gateway?” to diagnose a missing x-association-id header.
  • “How do I simulate a failed off-ramp in the sandbox?” to find the MSISDN rules.
  • “How do I configure and verify webhooks?” to reach Webhooks.

​
Security and limitations

  • Read-only: the server only searches and returns published docs. It cannot make API calls or change data.
  • Public content: it serves the published documentation. Do not paste secrets, tokens, or personal data into prompts.
  • Freshness: results reflect the current indexed site; unpublished changes are not searchable.
  • Rate limits: Mintlify applies per-user and per-site rate limits to the hosted server.