Skip to main content
Glama
F6-ZeppelinFellowship

Personal Knowledge-Base MCP Server

Personal Knowledge-Base MCP Server & Web App

F6-Zeppelin Fellowship — Project 3
A multi-tenant Model Context Protocol (MCP) server and web application enabling semantic search over personal document corpora backed by Qdrant vector search and FastMCP.


šŸš€ Overview

Static keyword search fails when notes, research papers, and technical documents use different wording for the same concepts. This project implements a protocol-level FastMCP server paired with a Qdrant Vector Database to enable context-aware semantic search over real-world documents.

The system supports dual modes of interaction:

  1. MCP Client Integration: Native tools callable from MCP-compliant clients like Claude Desktop or Claude Code.

  2. Multi-Tenant Web UI: A web dashboard providing isolated document management, uploading, and search capabilities per user.


Related MCP server: Tea Rags MCP

✨ Key Features

  • Protocol-Level Integration (FastMCP): Exposes structured MCP tools (search_notes, get_document, list_sources) for native AI agent invocation.

  • Multi-Tenant Isolation: Payload-level tenant isolation in Qdrant ensures document chunks and search results are strictly scoped per user.

  • Strict Relevance Cutoff: Rejects low-confidence vector matches below similarity thresholds to prevent low-relevance hallucination propagation.

  • Automated Ingestion Pipeline: Handles PDF, Markdown, and TXT parsing, dynamic chunking, and embedding generation.

  • Quantitative Retrieval Benchmarking: Hand-labeled evaluation suite tracking Mean Reciprocal Rank (MRR) and Precision@K across test queries.


šŸ› ļø Architecture & Tech Stack

Layer

Technology

Purpose

Protocol

FastMCP (Python)

Tool registry and JSON-RPC over STDIO / HTTP transport

Backend API

FastAPI

User authentication (JWT), file upload, REST search endpoints

Vector DB

Qdrant

HNSW similarity search with payload-based user isolation

Embeddings

sentence-transformers / OpenAI

Dense vectorization of document chunks

Frontend

React / Tailwind CSS

Web dashboard for uploading documents and testing queries


šŸ“Š Evaluation & Metrics

Metric

Target

Result

Precision@3

≄ 80%

TBD

MRR (Mean Reciprocal Rank)

≄ 0.85

TBD

Relevance Threshold

Cosine ≄ 0.72

Enforced


⚔ Quick Start

1. Environment Setup

cd backend
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env

2. Configure Claude Desktop (claude_desktop_config.json)

{
  "mcpServers": {
    "personal-kb": {
      "command": "python",
      "args": ["-m", "app.mcp_server.server"],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "QDRANT_API_KEY": "your-api-key"
      }
    }
  }
}

šŸ“‚ Repository Structure

KNOWLEDGE_BASE-MCP_SERVER/
ā”œā”€ā”€ backend/
│   ā”œā”€ā”€ app/
│   │   ā”œā”€ā”€ api/             # FastAPI REST endpoints (Auth, Documents, Search)
│   │   ā”œā”€ā”€ core/            # App configuration & JWT security settings
│   │   ā”œā”€ā”€ db/              # Qdrant vector database initialization & schemas
│   │   ā”œā”€ā”€ eval/            # Precision@K and MRR benchmark scripts
│   │   ā”œā”€ā”€ mcp_server/      # FastMCP server definition & tool implementations
│   │   └── services/        # Ingestion, embedding, & similarity search services
│   ā”œā”€ā”€ tests/               # Backend API and retrieval test suite
│   ā”œā”€ā”€ main.py              # Application entry point
│   ā”œā”€ā”€ requirements.txt     # Python backend dependencies
│   └── .env.example         # Template for environment variables
ā”œā”€ā”€ data/
│   └── sample_docs/         # Document corpus for local testing
ā”œā”€ā”€ docs/                    # Architecture diagrams & project documentation
ā”œā”€ā”€ frontend/                # React / Tailwind web application for multi-user management
│   └── src/
│       ā”œā”€ā”€ components/      # UI components (Uploaders, Search bar, Score badges)
│       ā”œā”€ā”€ context/         # Auth & Session state providers
│       ā”œā”€ā”€ pages/           # Document dashboard & Search playground
│       └── services/        # API client bindings
ā”œā”€ā”€ .gitignore               # Ignored files (venvs, keys, vector storage)
ā”œā”€ā”€ docker-compose.yml       # Local Qdrant & FastAPI orchestration
└── README.md                # Project documentation 
F
license - not found
-
quality - not tested
C
maintenance

Maintenance

–Maintainers
–Response time
–Release cycle
–Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • A
    license
    -
    quality
    F
    maintenance
    A Model Context Protocol (MCP) server that enables semantic search and retrieval of documentation using a vector database (Qdrant). This server allows you to add documentation from URLs or local files and then search through them using natural language queries.
    24
    135
    Apache 2.0
  • A
    license
    -
    quality
    A
    maintenance
    A high-performance MCP server for semantic search and codebase indexing using the Qdrant vector database. It features optimized embedding pipelines, AST-aware chunking, and git metadata enrichment for fast, privacy-focused local or remote search.
    685
    11
    MIT
  • A
    license
    -
    quality
    D
    maintenance
    MCP server for document ingestion and semantic search on Qdrant. Enables ingesting local documents, generating embeddings with OpenAI, and performing vector search with metadata filters.
    Apache 2.0
  • A
    license
    A
    quality
    D
    maintenance
    Local-first RAG indexing and semantic search MCP server. Enables document retrieval and context-aware queries using local embedding models.
    3
    9
    MIT

View all related MCP servers

Related MCP Connectors

  • Agentic search over your Dewey document collections from any MCP-compatible client.

  • Hosted MCP memory: save sessions/decisions once, search from Claude, Cursor, ChatGPT. EU-hosted FTS.

  • Agent-native MCP server over the public saagarpatel.dev corpus. Read-only, stateless.

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/F6-ZeppelinFellowship/Knowledge_Base-MCP_SERVER'

If you have feedback or need assistance with the MCP directory API, please join our Discord server