knowledge-mcp

A Model Context Protocol (MCP) server that provides scoped RAG workspaces ("Notebooks") backed by Qdrant and TEI.

Overview

This server enables an agent to:

1. Create named "Notebooks" (Qdrant Collections).
2. Ingest documents (PDF, Markdown, Text) into specific notebooks.
3. Query specific notebooks using vector search (RAG).
4. Synthesize findings across a notebook.

Designed to replicate the NotebookLM experience: clean, focused, bounded context.

Stack

• Language: Python 3.11+
• Framework: mcp SDK
• Vector DB: Qdrant
• Embeddings: Text Embeddings Inference (TEI) - BAAI/bge-base-en-v1.5

Tools

notebook.create

Creates a new isolated workspace (Qdrant Collection).

• name: string (e.g., "project-alpha")

notebook.add_source

Ingests a document into the notebook.

• notebook: string
• url: string (URL or local path)

notebook.query

Performs a semantic search/RAG generation against the notebook.

• notebook: string
• query: string

Configuration

Env vars:

• QDRANT_URL: URL to Qdrant (e.g., http://qdrant.openshift-gitops.svc:6333)
• TEI_URL: URL to TEI (e.g., http://text-embeddings.tei.svc.cluster.local:8080)