reverie-mcp-server
If you are the rightful owner of reverie-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
reverie-mcp-server is a knowledge base server designed for AI and developers, facilitating cross-project and cross-team collaboration through the Model Context Protocol (MCP).
reverie-mcp-server is a robust knowledge base server that leverages the Model Context Protocol (MCP) to support cross-project, cross-application, and multi-level permissions for knowledge storage and retrieval. It is designed to enhance collaboration between AI and developers by providing a structured and RAG (vector search) knowledge management system. The server allows for the efficient accumulation and reuse of best practices across different teams and projects. With its atomic Markdown knowledge units, reverie-mcp-server ensures easy maintenance and version control. The integration with Qdrant vector database enables semantic search capabilities, while the MCP Tool/Resource API facilitates knowledge operations accessible by both AI and users. The server also supports various permission levels, ensuring knowledge isolation and security.
Features
- Atomic Markdown Knowledge Units: Each knowledge point is an independent Markdown file, facilitating easy maintenance and version control.
- RAG Capability: Automatically syncs Markdown to Qdrant vector database for semantic search.
- Multi-project/Multi-app/Multi-team: Supports filtering and isolation by projectType, appId, owner, team, visibility, etc.
- MCP Tool/Resource API: Exposes list/search/add/delete knowledge operations via MCP protocol.
- Permission Levels: Supports private, team, and public knowledge isolation.
Tools
listKnowledge
List knowledge with multi-dimensional filtering and permission isolation.
searchKnowledge
Structured/semantic search (can be extended to RAG).
addKnowledge
Securely add new knowledge.
deleteKnowledge
Securely delete knowledge.