API docs

AI and MCP

Connect your AI assistant to these docs, so it writes WeSend integrations from the current API.

WeSend runs a Model Context Protocol (MCP) server. An AI assistant connected to it can read and search this documentation while it writes code, so it takes field names, limits and error codes from the current API instead of guessing them.

MCP URL

https://wesend.mn/mcp

The server only serves documentation. It needs no API key, cannot send messages and never sees your project.

Connect

claude mcp add --transport http wesend https://wesend.mn/mcp

Claude Code: run the command in your project folder.
Cursor: save it as .cursor/mcp.json in your project.
VS Code: save it as .vscode/mcp.json in your project.
In Claude, ChatGPT and other apps, add a custom connector (a remote MCP server) with the address above. There is nothing to sign in to.

Tools

Tool	What it does
list_docs	Lists the pages of this documentation and the endpoints each one covers.
read_doc	Returns one page in full: fields, limits, error codes and request examples.
search_docs	Searches by error code, field or topic and returns the sections that match.

Every tool takes lang (en or mn). Request examples come as curl, node or python, chosen with examples.

Use it

Once connected, ask for what you need in plain words. The assistant looks the details up itself.

“Add phone verification with WeSend to the sign-up form.”
“Send the order confirmation by SMS, and retry without sending twice if the request times out.”
“What does recipient_rate_limited mean, and how should I handle it?”

Do not paste your API key into the chat. Keep it in the WESEND_API_KEY environment variable: every example the assistant reads takes it from there.