kk-knowledge-agent

轻量级知识库 Agent 项目，支持知识库管理、文本/txt 上传、分块向量化、语义检索、流式返回，并通过 MCP Server 暴露给 Agent 调用。

Architecture

Frontend
   |
   | HTTP
   v
Backend API
   |
   +--> SQLite
   +--> ChromaDB
   +--> sentence-transformers

Agent / Claude Code / OpenClaw / Hermas Agent
   |
   | MCP Tool Call
   v
MCP Server
   |
   | HTTP
   v
Backend API

Repository Layout

backend/       FastAPI API service
frontend/      React + Vite demo app
mcp-server/    MCP adapter server
examples/      Demo txt documents
AGENTS.md      Phased development guide for Codex/Agent

Demo Screenshots

Frontend knowledge base and document workflow:

Codex MCP knowledge base tool call:

Local Setup

Backend

cd backend
python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt
copy .env.example .env
uvicorn app.main:app --reload --port 8000

API docs:

http://localhost:8000/docs

Frontend

cd frontend
npm install
npm run dev

Demo app:

http://localhost:5173

MCP Server

cd mcp-server
python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt
copy .env.example .env
python server.py

MCP config example:

{
  "mcpServers": {
    "kk-knowledge": {
      "command": "python",
      "args": ["mcp-server/server.py"]
    }
  }
}

MVP Scope

• Knowledge base CRUD with pagination.
• Direct text document upload.
• .txt file upload.
• Chunking and local multilingual embedding.
• ChromaDB semantic retrieval.
• SSE streaming search.
• MCP tool: search_knowledge_base.
• Basic tests with pytest.

Development Order

Follow AGENTS.md for the phased implementation plan.

Local Installation

See docs/LOCAL_INSTALLATION.md for detailed Backend, Frontend, and MCP Server setup.

Future Improvements

• Add Docker Compose for one-command local deployment.
• Replace SQLite with PostgreSQL for production usage.
• Replace ChromaDB with Qdrant or Milvus for larger vector workloads.
• Add PDF/DOCX upload.
• Add BM25 + vector hybrid retrieval and rerank support.
• Add answer generation with citation highlights.