1C MCP Server with a FastAPI proxy
1C MCP Server
A 1C extension with an MCP HTTP service plus a Python FastAPI proxy that adds stdio transport, OAuth2 and Docker deployment
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- The proxy stores the 1C user's login and password in environment variables
- The extension's HTTP service gives access to the published infobase's data and metadata
Install
Manual install
cd src && python -m venv .venv && . .venv/bin/activate && pip install -r py_server/requirements.txt && python -m py_serverRuns the Python proxy locally with fastapi, uvicorn, httpx, pydantic and mcp as dependencies.
This is third-party code. Review the repository files before installing.
What it does
The project offers two ways to connect an MCP client to 1C: directly to the extension's published mcp_APIBackend HTTP service, or through a Python FastAPI proxy that adds stdio transport for clients without HTTP support, OAuth2 authorization, and pass-through authentication into the 1C backend. The extension ships as a ready .cfe file, and the proxy runs locally via venv or in Docker with environment variables for the 1C infobase address, username and password. The repository includes example configs for Claude Code, Claude Desktop, Cursor and Kilo Code.
Who it is for. For 1C developers who need a ready MCP server with flexible transport and centralized authorization, not just a direct HTTP connection.
Good fit when
- You need stdio transport for an MCP client that cannot talk HTTP directly
- You need centralized OAuth2 authorization in front of the 1C backend
- Docker deployment with a ready compose file is more convenient for you
Not a fit when
- A direct HTTP connection to the extension without a proxy is enough
- You are not ready to publish the mcp_APIBackend HTTP service in your infobase
Example request
Connect to 1C through the MCP proxy and show the metadata of the Nomenclature catalogLimitations
The extension has to be installed in the infobase and its HTTP service published; the README does not walk through importing the .cfe step by step. The 1C user password is passed to the proxy via environment variables. The project is small, with one author, no releases and no stars.
How to disable. Stop or remove the Python proxy, disable the mcp_APIBackend extension in the 1C infobase, and remove the server from your MCP client settings.
MCP
- Transport
- http, stdio
- Authentication
- API key
| Environment variables | |
|---|---|
| MCP_ONEC_URL required | URL of the published 1C infobase. |
| MCP_ONEC_USERNAME required | 1C user login. |
| MCP_ONEC_PASSWORD required, secret | 1C user password. |
| MCP_AUTH_MODE | Proxy authorization mode: none or oauth2. |
Security check
- The proxy stores the 1C user's login and password in environment variables
- The extension's HTTP service gives access to the published infobase's data and metadata
README in short
The README describes the combination of a 1C extension and a Python FastAPI proxy with fastapi, uvicorn, httpx, pydantic and mcp as dependencies. It gives three ways to run it: installing the extension and publishing the HTTP service, running the proxy locally via venv, and running it in Docker with .env.docker.example. It lists the project layout: build with the .cfe, src with the sources, and mcp_client_settings with example client configs.
FAQ
Is the Python proxy required?
No, the README lists it as the recommended option for stdio transport and centralized authorization, but you can also connect directly to the extension's HTTP service.
How do I enable OAuth2?
Set the MCP_AUTH_MODE variable to oauth2 in the proxy configuration.
Related
A skills library that gives coding agents a development process: brainstorming, planning, TDD, subagents and code review
Skills for real engineers by Matt Pocock
Skills For Real Engineers
Small composable skills for engineering with agents: plan grilling, TDD, bug diagnosis, code review and architecture
GitHub toolkit for spec-driven development: the specify CLI adds agent commands and skills to a project, from principles to implementation
Reference MCP servers
Model Context Protocol servers
Official reference MCP servers: Filesystem, Fetch, Git, Memory, Sequential Thinking, Time and Everything