1C MCP: search over a configuration's structure and code
1C MCP
A stdio MCP server that indexes a 1C structure XML export and code into SQLite and gives an agent search over objects and procedures
Low risk
We rate an entry low when it mostly gives the agent instructions and reference material.
Why this level
- The server only reads the SQLite index and does not write to the 1C database
Install
Manual install
docker compose run --rm --build index-structure && docker compose run --rm --build index-docs && docker compose run --rm -T mcpBuilds the structure and documentation index and immediately starts the MCP server, meant for a first run on a new machine.
This is third-party code. Review the repository files before installing.
What it does
The project does not store the 1C configuration itself, it builds an index over it. From a structure XML export it builds SQLite tables of metadata objects, attributes and type links. Separately, you can index a folder of exported 1C code: the indexer finds procedure and function declarations and endings in both Russian and English syntax and pulls them into searchable records. A third layer is documentation: the indexer reads markdown, text, html and PDF files from a docs folder and makes them searchable through the search_docs tool. The MCP server runs over stdio and gives the agent about a dozen search tools across all three layers at once.
Who it is for. For 1C developers who want to give an agent fast search over a specific configuration's structure and code instead of manually grepping the export.
Good fit when
- You have a configuration structure XML export and need to search objects and attributes quickly
- You have a 1C code export and need to search procedures and functions by name or content
- You want reference documentation next to the code index for search in one place
Not a fit when
- You have neither a structure XML export nor a code export, so there is nothing to index
- You need access to a live 1C database, not a static index of an export
- You are not ready to keep Docker Desktop running for indexing and serving
Example request
Find the procedure that recalculates balances in the Товары register and show its source codeLimitations
The server only works with prepared exports, not with a live 1C database directly. The repository ships a few 1C-related PDF books alongside the code as a documentation-indexing example, and their origin and rights are not stated, so it is worth replacing that folder with your own material for a real project. Indexing without rebuilding the image after a Dockerfile change will not pick up new dependencies like pypdf.
How to disable. Stop the containers with docker compose down; the MCP server only runs when a client launches it and does not linger in the background on its own.
MCP
- Transport
- stdio
- Authentication
- not required
| Environment variables | |
|---|---|
| CODE_EXPORT_DIR | Host path to the folder with exported 1C code to index |
Security check
- The server only reads the SQLite index and does not write to the 1C database
README in short
The README explains that the project stores an index rather than the configuration itself, lists the SQLite tables and indexer modules, shows docker compose commands to run the MCP server and to index structure, code and documentation separately, gives an MCP client config example both through docker compose run and a direct python3 mcp/server.py launch, and describes the formats the code and documentation indexers understand.
FAQ
Do I need to reindex on every configuration change?
Yes, if the structure XML export changed or a new code export appeared, the README recommends rebuilding the relevant index.
Can it work without Docker?
The local scripts parser/indexer.py, code_indexer.py, docs_indexer.py and search/search.py remain as helpers, but the README recommends Docker as the main way to run it.
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