knowledge-mcp-server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@knowledge-mcp-server search my indexed documents for Python tutorials"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
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/activateRun tests:
uv run pytestRelated MCP server: pdf-knowledge-mcp
Use It With Your Own Knowledge Set
Start from a clean data state (optional but recommended):
rm -f data/docs/* data/vectorstore/*
echo '{}' > data/metadata.jsonIndex 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.texQuery 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 8000Open 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-v2Chunking: 800 words, 100 overlap
HTTP port:
8000
Docs
docs/MCP_SETUP.md- MCP registration and troubleshootingdocs/ARCHITECTURE.md- architecture notes
License
MIT (see LICENSE)
Available Tools
4 toolsadd_documentB
Add a markdown document to the knowledge base
Inputs: markdown (required), title, repo.
index_urlB
Fetch and index content from a URL into the knowledge base
Inputs: url (required).
list_documentsA
List all documents in the knowledge base
Inputs: no named parameters.
query_knowledgeA
Search the knowledge base using semantic search. Returns relevant document chunks with similarity scores.
Inputs: query (required), top_k.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool has a clearly distinct purpose: adding documents, indexing URLs, listing documents, and querying. No overlap or ambiguity.
All tool names follow a consistent verb_noun pattern in snake_case (add_document, index_url, list_documents, query_knowledge).
With 4 tools, the server is appropriately scoped for a knowledge base manager, covering essential operations without excess.
The toolset covers adding (two methods), listing, and searching the knowledge base, but lacks update or delete operations, which is a minor gap.
Related MCP Connectors
Self-hosted AI workspace: semantic RAG document search, long-term memory, and AI chat via MCP.
An MCP server that gives your AI access to the source code and docs of all public github repos
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
Cloud or self-hosted knowledge for AI agents: hybrid search, reranking, GraphRAG, scoped MCP tools.
Related MCP Servers
- AlicenseAqualityDmaintenanceRAG MCP server for PDFs, YouTube, GitHub repos, and Discord exports. Index documents and query with citations via LangChain and Chroma.533 PyPI2MIT
- FlicenseAqualityDmaintenanceA local RAG MCP server for PDF development experience, enabling document ingestion, semantic search, and Q\&A with source citations using TF-IDF and cosine similarity.3-
- AlicenseBqualityBmaintenanceLocal end-to-end RAG system for agentic code editors, exposing retrieval-augmented generation via MCP to any compatible client.331MIT
- AlicenseNot gradedqualityDmaintenanceA local RAG MCP server that enables AI tools like Claude to search indexed codebases and documentation using vector search with Ollama models.Apache 2.0