1C AI MCP: 1C:Naparnik API bridge
1C AI MCP
A FastMCP Docker server with 12 tools for BSL code analysis and documentation search through the 1С:Naparnik API (code.1c.ai)
- 9 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Sends BSL code snippets and questions to the external code.1c.ai service
- Holds an access token in environment variables or Docker Secrets
Install
- Works in
- Services
Manual install
docker run -d --name 1c-ai-mcp -p 8007:8007 -e ONEC_AI_TOKEN="ваш-токен" desko77/1c-ai-mcp:latestThe ready Docker Hub image, the recommended method.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Sends BSL code snippets and questions to the external code.1c.ai service
- Holds an access token in environment variables or Docker Secrets
This is third-party code. Review the repository files before installing.
What it does
The server wraps the 1С:Naparnik neural network API in a set of MCP tools: code checks for syntax, logic and performance, code review against ITS standards, rewriting and instruction-driven modification, syntax explanations, and free-form questions with conversation context preserved across calls. A separate group of tools searches the ITS knowledge base and version-aware platform documentation, diffs documentation between versions, and searches application configuration docs like ERP or payroll. It has two call modes: standard builds text prompts, while direct calls specific upstream 1С:Naparnik tools by name with an automatic fallback to standard on failure.
Who it is for. For 1C developers with a 1С:Naparnik token who want to call that neural network right from Cursor or Claude Code as an MCP tool.
Good fit when
- You need a BSL code check or review against ITS standards right from the IDE
- You need quick search over platform or application configuration docs for a specific version
- You need to compare platform behavior between two versions
Not a fit when
- There is no 1С:Naparnik token: it is issued after registration at code.1c.ai
- The code or questions contain confidential data that cannot be sent to an external service
- You need an official vendor server: this is an independent community wrapper over the API
Example request
Review this document-posting handler against ITS standards and suggest improvementsMCP
- Transport
- http, sse
- Authentication
- API key
Environment variables
ONEC_AI_TOKENrequired, secret- 1С:Naparnik API token, required unless ONEC_AI_TOKEN_FILE is set.
ONEC_AI_TOKEN_FILEsecret- Path to a file holding the token, for use with Docker Secrets.
ONEC_CONFIG_NAME- Application configuration name for the config_help tool, for example ERP or payroll.
Limitations
A 1С:Naparnik token is required: according to code.1c.ai, the key is free until 1 February 2027. The project is based on the idea of the now-closed comol/1c-code-checker repository, with API fixes borrowed from SteelMorgan/spring-mcp-1c-copilot, meaning this is an independent community wrapper, not a vendor server. The direct mode depends on specific 1С:Naparnik upstream tool names and may need updates if those change on the service side.
How to disable. Stop the container with docker compose down or docker stop 1c-ai-mcp and remove the server from your MCP client settings.
FAQ
How does direct mode differ from standard?
Standard builds text prompts and relies on model choice, direct calls specific 1С:Naparnik upstream tools by name for more precise results, falling back to standard on failure.
Can the token be passed without an environment variable in plain sight?
Yes, through ONEC_AI_TOKEN_FILE and Docker Secrets for production environments.
How is the MCP server for the 1C:Naparnik API different from 1C:Naparnik itself?
A Docker server on FastMCP wraps the 1C:Naparnik AI API (code.1c.ai) in 12 MCP tools: BSL code checks for syntax, logic and performance, code review against ITS standards, code rewriting and documentation search.
How does direct mode differ from standard and how do I pass the token?
The token is passed as the ONEC_AI_TOKEN environment variable when starting the container, as in the install block command. The direct and standard mode differences are described in the project README. It connects to Cursor and Claude Code.
Discussion
No questions or reviews yet.
Related
All 1C:Enterprise tools1C AI Assistant MCP proxy
1c-ai-mcp
An MCP server that forwards questions, code review and code explanation requests to the 1C AI Assistant cloud service
1C:Naparnik MCP bridge in Go
1c-copilot-mcp
A Go binary with 5 MCP tools for the 1С:Naparnik neural network that reads 1C code straight from a file path instead of through the model's context
MCP server for 1C:Naparnik
Spring Boot MCP Server для 1С:Напарник
Spring Boot and Kotlin MCP server that connects an agent to the 1C:Naparnik API for 1C questions and code checks
An MCP server that exposes the cloud 1C:Napparnik (1C.ai) assistant and a local BSL Language Server to agents in Cursor and other IDEs
1C Development Standards MCP
MCP-сервер стандартов разработки 1С:Предприятие 8
An MCP server exposing 1C:Enterprise 8 development standards for agent-driven search and code review
Five independent Docker-based MCP servers: a Neo4j metadata and call graph, static BSL analysis, platform syntax help, a query builder and 1C coding standards
README in short
The README describes a FastMCP Docker server with 12 tools grouped into code analysis and documentation search, gives two run options via a ready Docker Hub image or building from source, an environment-variable configuration table, Docker Secrets for the token, an architecture diagram from the MCP client to the code.1c.ai API, an explanation of standard and direct modes with specific upstream tool names, and the SSE parser's structure with three supported response formats. It credits the sources of the idea and API algorithms. MIT license.