1C OData MCP with a JSON cache
MCP сервер для работы с OData 1C
A Python MCP server over 1C OData: reads and changes catalogs and documents, saving everything to JSON files for local analysis without repeat queries
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- The create_1c_entity, update_1c_entity and delete_1c_document tools directly create, change and delete documents and catalog items in 1C
- Basic Auth credentials are stored in plain text in .env, and exports are saved to disk locally
Install
Manual install
uv syncInstall dependencies.
This is third-party code. Review the repository files before installing.
What it does
The server connects to 1C's OData API with Basic Auth and gives an agent tools for reading metadata and data with automatic saves to JSON files, plus CRUD operations: creating and updating entities, deleting documents, and a separate builder for goods-and-services documents with business rules. The core architectural idea is that data-fetching functions return only a status while the export itself stays in a file: filter_local_data, get_file_structure and count_local_data then work against that file locally, without repeat calls to 1C, saving model context. It includes smart file lookup by partial name and cache management for both metadata and data.
Who it is for. For 1C developers and analysts who want an agent that reads and changes 1C data over OData, with exports preserved between sessions.
Good fit when
- You need to export and locally analyze a large volume of 1C data without repeat server queries
- You need CRUD operations on catalogs and documents over OData from an agent
- You need to create goods-and-services documents based on other documents, respecting business rules
Not a fit when
- The database's OData interface is not published, or publishing it is unsafe for the environment
- You need operation without saving data to disk: the server's whole architecture is built on local JSON-file caching
- You need authentication stronger than Basic Auth: the README doesn't describe other methods
Example request
Export the Products catalog, show the resulting file's structure, and sort the first 10 items by priceLimitations
Authentication to 1C is Basic Auth only, with the login and password stored in plain text in .env. All exported data is saved to JSON files on disk, including potentially personal data from catalogs and documents, so access to the cache directory needs to be restricted separately. The create_1c_document and update_1c_document functions are marked deprecated in favor of create_1c_entity and update_1c_entity. Python 3.12 or newer and a configured 1C OData API are required.
How to disable. Stop the uv run main.py process and remove the server from your MCP client configuration; cache files can be deleted manually from the working directory.
MCP
- Transport
- sse
- Authentication
- API key
| Environment variables | |
|---|---|
| ODATA_BASE_URL required | Address of the 1C database's OData API. |
| ODATA_USERNAME required | Username for Basic Auth to 1C. |
| ODATA_PASSWORD required, secret | Password for Basic Auth to 1C. |
Security check
- The create_1c_entity, update_1c_entity and delete_1c_document tools directly create, change and delete documents and catalog items in 1C
- Basic Auth credentials are stored in plain text in .env, and exports are saved to disk locally
README in short
The README lists the server's features: metadata, data queries, local filtering, CRUD operations, special goods-and-services documents, Basic Auth authentication, and a JSON architecture with a rationale for saving LLM context. It gives a quick start via uv sync and .env, an SSE connection example, a full tool list by group, code examples for creating and updating entities, a description of smart partial-name file lookup, and a link to architecture.md for details.
FAQ
Is data fetched from 1C fresh every time?
No, fetch_1c_data saves the export to a JSON file once, and further analysis via filter_local_data and count_local_data works against that file locally.
Can a document be deleted through this server?
Yes, the delete_1c_document tool removes a document from 1C, which is a write operation, so access to the server should be restricted.
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