RAG MCP server for 1C configuration structure
mcp-1c-v1
A container set: 1C configuration structure export, a Qdrant vector database and an MCP server that finds objects by meaning
Install
chmod +x start.sh && chmod +x stop.sh
./start.shRun inside the cloned repository; alternative is docker-compose up --build.
This is third-party code. Review the repository files before installing.
What it does
The project is a set of services. An external data processor exports the configuration structure from 1C as markdown files, a Streamlit loader puts the archive into Qdrant, an embedding service builds vectors, and the MCP server answers the agent's questions about metadata. Search runs over two vectors, object name and synonym, merged with RRF. The x-collection-name header lets one server handle different configurations.
Who it is for. For 1C developers who want to find configuration objects by natural-language description.
Good fit when
- You need to find a metadata object without knowing its exact name
- You want one structure server for several configurations
- You want a self-hosted RAG for 1C
Not a fit when
- You need search over module code rather than structure
- You cannot run Docker with several services
Example request
Find the documents responsible for customer product returnsLimitations
Only the configuration structure is indexed, not code. After changes you must export and upload the archive again. The project has not been updated for a long time. Requires Docker Compose; image builds download dependencies and the embedding model.
How to disable. Run ./stop.sh or docker-compose down and remove the server from the client settings.
MCP
- Transport
- http
- Authentication
- not required
| Environment variables | |
|---|---|
| EMBEDDING_SERVICE_URL | Embedding service URL |
| QDRANT_HOST | Qdrant host |
| QDRANT_PORT | Qdrant port |
| COLLECTION_NAME | Default collection, 1c_rag |
Security check
- Runs several network services without built-in authentication
- Stores the configuration structure in a vector database
README in short
The README lists the project's services, describes multi-vector search and starting all containers via a script or docker-compose. It gives URLs for the loader, embedding service, Qdrant, MCP server and MCP Inspector. It walks through the flow from structure export to search, with VS Code and Cursor configs and environment variables. There is a video and an article by the author.
FAQ
Why does VS Code Copilot fail to connect?
The author notes Copilot worked for him only over SSE, so use the /mcp/sse URL.
How do I keep several configurations?
Upload each into its own Qdrant collection and set its name in the x-collection-name header in project settings.
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
Reference MCP servers
Model Context Protocol servers
Official reference MCP servers: Filesystem, Fetch, Git, Memory, Sequential Thinking, Time and Everything
Up-to-date, version-specific library docs and code examples in your agent's context, via MCP or a CLI plus skill