Evolution Managed RAG MCP server
Agentic RAG Telegram Bot + MCP Managed RAG
A Cloud.ru teaching MCP server on top of Evolution Managed RAG: hands the agent context from a knowledge base, with a sample GigaChat Telegram bot included
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Uses Cloud.ru service account keys with access to the project's knowledge base
- The example Telegram bot sends messages to chat users on the bot's behalf
Install
Manual install
docker compose up --buildBrings up both services from the README. Set the bot's and MCP server's environment variables first.
This is third-party code. Review the repository files before installing.
What it does
The project has two services. mcp-managed-rag runs an SSE MCP server with one tool, request_to_rag: it queries an already configured Evolution Managed RAG knowledge base on Cloud.ru, formats the matching documents into context, and returns it to the agent. bot-managed-rag shows how a LangGraph Telegram bot on GigaChat uses that MCP tool: a user messages the bot, a ReAct agent calls the remote MCP tool, and the answer streams back into Telegram by editing the message as it generates.
Who it is for. For people who already have a Cloud.ru Evolution Managed RAG knowledge base and need an example MCP wrapper and Telegram bot on top of it.
Good fit when
- You already have a knowledge base in Evolution Managed RAG and need MCP access to it for an agent
- You want a working example of a RAG Telegram bot on GigaChat and LangGraph
- You need a sample SSE MCP server to deploy your own tool on Cloud.ru
Not a fit when
- You have no knowledge base in Evolution Managed RAG: without it there is nothing for the server to query
- You need a ready public server: both services must be deployed and configured yourself
Example request
Find the return policy answer in the knowledge base and summarize it in your own wordsLimitations
A Cloud.ru lab teaching project that needs an already created Evolution Managed RAG project and knowledge base with service account keys, plus GigaChat API access and a Telegram bot token for the bot example. MCP transport is currently SSE only. It has had no pushes or updates for over a year.
How to disable. Stop the services with docker compose down and remove the MCP server from your client config.
MCP
- Transport
- sse
- Authentication
- API key
| Environment variables | |
|---|---|
| EVOLUTION_PROJECT_ID required | The Evolution project ID in Cloud.ru. |
| KNOWLEDGE_BASE_ID required | The Managed RAG knowledge base ID. |
| KNOWLEDGE_BASE_VERSION_ID required | The knowledge base version ID. |
| SERVICE_ACCOUNT_KEY_ID secret | The Cloud.ru service account key ID. |
| SERVICE_ACCOUNT_KEY_SECRET secret | The Cloud.ru service account key secret. |
| GIGACHAT_CREDENTIALS required, secret | The GigaChat authorization key for the Telegram bot example. |
| TELEGRAM_BOT_TOKEN required, secret | The Telegram bot token for the bot-managed-rag example. |
Security check
- Uses Cloud.ru service account keys with access to the project's knowledge base
- The example Telegram bot sends messages to chat users on the bot's behalf
README in short
The Russian README describes the two-service architecture and the request path from Telegram through Managed RAG and back, lists all environment variables for the bot and the MCP server separately, and gives the docker compose up --build command for a quick start.
FAQ
Do I have to deploy the Telegram bot?
No, bot-managed-rag is just one example client. The mcp-managed-rag server can be connected directly to any SSE MCP client.
Do I need a separate knowledge base?
Yes, the server only reads an existing Evolution Managed RAG knowledge base; creating and filling it happens separately in Cloud.ru.
Related
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
An open source coding agent for the terminal and desktop with build and plan modes
Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent