fast-bitrix24-mcp
MCP server for Bitrix24 CRM analytics: deals, leads, contacts, tasks, manager activity and deals at risk
- 2 ★ on the repository
- Last commit
Extended access
Why this level
- Reads CRM data, including client personal data
- The webhook may carry broader rights than reading requires
- Exports data to files on disk
Install
- Works in
- Services
Manual install
cp .env.example .env
uv add fast-bitrix24-mcp
uv run main.pyCreate main.py yourself: import mcp from fast_bitrix24_mcp.main and call mcp.run(transport="http", host="0.0.0.0", port=8000). .env needs WEBHOOK and AUTH_TOKEN.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Reads CRM data, including client personal data
- The webhook may carry broader rights than reading requires
- Exports data to files on disk
This is third-party code. Review the repository files before installing.
What it does
Built on the fast-bitrix24 library, the server gives the agent filtered lists of deals, leads, contacts, companies, users and tasks, human-readable funnel stages and decoded custom fields, including list fields. Ready report tools cover the sales funnel for a period, top clients by deal sum, a daily summary, deals at risk, each manager's activity and managers whose activity dropped. Large selections are first exported to a JSON file and then computed by a separate analysis tool: sum, count, average, grouping. A main prompt explains the workflow to the agent.
Who it is for. For sales heads and analysts who want to query Bitrix24 CRM data in plain language.
Good fit when
- You need a deal sum by a custom field value
- You need managers with overdue tasks or dropped activity
- You need a daily summary of leads, deals and calls
Not a fit when
- You need to create or change deals: the tools focus on reading and analysis
- You need a local stdio server without running HTTP
Example request
What is the total of deals where the delivery floor field equals basement? And which managers have overdue tasks?MCP
- Transport
- http
- Authentication
- API key
Environment variables
WEBHOOKrequired, secret- Bitrix24 inbound webhook
AUTH_TOKENrequired, secret- Token the client sends in the Authorization header
Limitations
Requires a Bitrix24 inbound webhook. The server runs as an HTTP service from your own Python script and accepts only requests with AUTH_TOKEN. Criteria for deals at risk and dropped activity are fixed and may not match your rules. Exports write CRM data files to the server disk. Python 3.12, last changes in late 2025.
How to disable. Stop the server process, remove the connection from the MCP client and disable the webhook in Bitrix24 settings. Delete exported JSON files manually.
FAQ
Does it understand field names instead of UF_CRM codes?
Yes, get_all_info_fields returns field names and values, including list fields, and the agent matches them to the question.
What counts as a deal at risk?
A deal whose status has not changed for more than five business days or has no calls, comments or tasks for more than three days.
Discussion
No questions or reviews yet.
Related
All Bitrix24 toolsBitrix24 MCP server with sales analytics
Bitrix24 MCP Server
Bitrix24 MCP server in TypeScript with a broad set of CRM tools plus a dedicated group for monitoring manager activity and performance
Bitrix24 MCP Server (gunnit)
Bitrix24 MCP Server
A TypeScript MCP server for Bitrix24 CRM: contacts, companies, leads, deals, tasks and sales rep performance reports
An MCP server for Bitrix24 CRM contacts and deals via an inbound webhook: search, lists and deal stage changes
Bitrix24 MCP by Bit2Beat
Bitrix24 MCP Server
Bitrix24 MCP server over an inbound webhook: CRM, tasks, disk, calendar, chat, business processes and config migration between portals
A read and write MCP server for the Bitrix24 REST API: CRM, tasks, scrum, calendar, disk, messaging and portal events
A read-only MCP server that reads Bitrix24 deals, contacts and companies with SELECT queries against the database, bypassing the REST API
README in short
The Russian README lists tools by entity and shows plain-language example questions. Install uses uv, the server starts from a small script in HTTP mode, and requests are authorized by a token. It includes a LangChain and LangGraph connection example, explains the deals at risk criteria and ships the main prompt with tool descriptions.