knowledge-mcp-server

Small, forkable RAG project with an MCP server (for VS Code/Copilot) and a minimal web UI.

Quick Start

Prerequisite: install uv from https://docs.astral.sh/uv/

uv sync
source .venv/bin/activate

Run tests:

uv run pytest

Related MCP server: pdf-knowledge-mcp

Use It With Your Own Knowledge Set

1. Start from a clean data state (optional but recommended):

rm -f data/docs/* data/vectorstore/*
echo '{}' > data/metadata.json

2. Index content:

uv run python -m knowledge_mcp_server.cli index --url https://example.com
# or --pdf /path/to/file.pdf
# or --tex /path/to/file.tex

3. Query via MCP or HTTP/web UI.

Register MCP Server In VS Code

code --add-mcp '{"name":"knowledge-mcp-server","command":"uv","args":["run","python","-m","knowledge_mcp_server.mcp_stdio_server"],"cwd":"'$(pwd)'"}'

Run Web UI

uv run uvicorn knowledge_mcp_server.mcp_server:app --reload --port 8000

Open http://127.0.0.1:8000/.

Project Layout

• Code: src/knowledge_mcp_server/

• Tests: tests/

• Runtime data: data/docs/, data/vectorstore/, data/metadata.json

Defaults

• Embeddings: all-MiniLM-L6-v2

• Chunking: 800 words, 100 overlap

• HTTP port: 8000

Docs

• docs/MCP_SETUP.md - MCP registration and troubleshooting

• docs/ARCHITECTURE.md - architecture notes

License

MIT (see LICENSE)