reverie-mcp-server

reverie-mcp-server

3.2

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

A cross-project, cross-team, and best-practice knowledge base server for AI and developers

Overview

reverie-mcp-server is a knowledge base server based on the Model Context Protocol (MCP). It supports cross-project, cross-application, multi-level permissions, and both structured and RAG (vector search) knowledge storage and retrieval. It enables AI and developers to efficiently collaborate, accumulate, and reuse best practices.

Features

  • Atomic Markdown Knowledge Units: Each knowledge point is an independent Markdown file, easy for manual maintenance and version control
  • RAG Capability: Automatically sync 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, accessible by AI and users
  • Permission Levels: Supports private, team, and public knowledge isolation
  • Best Practice Accumulation: Extensible for auto-summarization, recommendation, and completion tools

Quick Start

Requirements

  • Node.js >= 18
  • Qdrant vector database (local or cloud)
  • Ollama embedding API (or custom embedding service)

Install dependencies

npm install

Start MCP Server

npm run build
npx reverie-mcp-server 

Directory Structure

knowledge/
  items/                # Markdown knowledge units
src/
  knowledge/           # Structured knowledge and permission logic
  rag/                 # RAG/vector search logic
  index.ts             # MCP Server entry point

Main APIs (MCP Tool)

  • 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

Markdown Knowledge Unit Example

---
id: xxx-001
title: Feature Implementation
summary: Brief principle/API
projectTypes: [miniapp, taro]
appIds: [app1, app2]
visibility: public
owner: team-abc
tags: [navigation, jump]
createdAt: 2024-06-01T12:00:00Z
updatedAt: 2024-06-01T12:00:00Z
demoUrl: https://demo.com
docUrl: https://docs.com
---

## Effect Overview
...

## Principle/API
...

## Code Example
```js
// Example code

## Development & Extension
- All core logic is modular, easy to extend with more Tool/Resource
- Supports custom embedding, RAG backend, and permission models
- Recommended to manage Markdown knowledge files with Git for team collaboration

## Contribution
PRs, issues, and discussions are welcome!

## License
GNU General Public License v3.0 (GPLv3)

This project is licensed under the GNU GPL v3. See the LICENSE file for details.