Beever Atlas
Free开源团队知识库，内建 MCP 服务器。Neo4j 类型化知识图谱 + Weaviate 向量数据库，支持 Slack/Discord/Teams 等对话导入，通过 LiteLLM 自选 LLM。
FreeFree tier
About Beever Atlas
Beever Atlas is an open-source team knowledge base that automatically turns conversations from Slack, Discord, Microsoft Teams, and Mattermost into a self-maintaining wiki. It uses a 6-stage ingestion pipeline to extract atomic facts and relationships, stored in Neo4j graph database and Weaviate vector database. Users can ask natural language questions and receive cited answers through a dashboard or via MCP into Claude Code and Cursor. The system offers multi-platform connectivity, scheduled message sync, auto-maintained wiki pages per channel, and a QA agent with smart routing between semantic and graph search.
Key Features
Multi-Platform: Connect Slack, Discord, Teams, Mattermost, or file imports with one bot per workspace
Message Sync: Pull channel history on demand or on schedule, resumable and rate-limit aware
Memory Ingestion: 6-stage ADK pipeline distills messages into atomic facts, entities, and relationships
LLM Wiki: Auto-maintained wiki per channel with overview, topics, people, decisions, and citations
QA Agent: Streams cited answers over SSE with a smart router that picks semantic or graph search per question
MCP Server: Plug Claude Code/Cursor into the knowledge base with 28 tools and per-agent authentication
Graph Store: Neo4j typed knowledge graph linking people, decisions, and projects across channels
Vector Store: Weaviate vector database for fast hybrid search
Pros & Cons
Pros
- Fully open-source with no licensing costs
- Automatic wiki generation reduces manual documentation effort
- Supports multiple chat platforms out of the box
- Cited answers ensure transparency and traceability to source messages
- MCP integration allows direct use with popular AI coding tools
- Dual memory system (graph + vector) provides accurate and contextual retrieval
- Resumable and rate-limit aware message sync for reliability
Cons
- Requires self-hosting and setup of multiple services (backend, bot, frontend) and databases
- Dependent on LLM choice via LiteLLM, which may incur API costs
- Limited to supported chat platforms (Slack, Discord, Teams, Mattermost)
- Ingestion pipeline complexity may require technical expertise to configure and tune
- No pre-built hosted version; users must deploy from source or Docker
Best For
Turn team Slack/Discord/Teams chats into a self-maintaining wiki automaticallyAsk natural language questions about past decisions and get cited answersIntegrate team knowledge into Claude Code and Cursor via MCP serverMaintain a living knowledge base from ongoing conversations without manual effortTrack who said what, decisions made, and project relationships across channels
FAQ
What chat platforms does Beever Atlas support?
It supports Slack, Discord, Microsoft Teams, Mattermost, and file imports.
How does Beever Atlas store knowledge?
It uses a dual-memory system: Neo4j graph database for entities and relationships, and Weaviate vector database for semantic search.
Can I use Beever Atlas with Claude or Cursor?
Yes, it provides an MCP server with 28 tools that can be plugged into Claude Code and Cursor with per-agent authentication.
Is Beever Atlas free?
Yes, it is open-source and free to use. You need to self-host it.
Does Beever Atlas automatically update the wiki?
Yes, the LLM Wiki is auto-maintained per channel with overview, topics, people, decisions, and citations as new messages are ingested.