YouTube Knowledge Base

Provides a searchable knowledge base built from YouTube transcripts with hybrid search and timestamped moment links.

python

2

GitHub Stars

python

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "miandari-youtube-knowledge-base-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/youtube-knowledge-base-mcp",
        "run",
        "youtube-kb"
      ],
      "env": {
        "YOUTUBE_KB_DATA_DIR": "/custom/data/path"
      }
    }
  }
}

You build a searchable knowledge base from your YouTube videos and access it through your favorite AI assistant. This MCP server ingests transcripts, enables hybrid search, and provides timestamped links to exact moments, so your memory is augmented and easy to navigate.

How to use

You can connect this MCP server to an AI assistant client to add videos to your knowledge base or search across your saved content. Use the Claude Desktop path or a Python workflow to ingest and query videos. When you add a video, the system extracts its transcript, associates it with tags and notes, and stores embeddings from your chosen provider.

How to install

Prerequisites: Python 3.10+, uv package manager. You will also need at least one embedding/provider key (Voyage, OpenAI, or a local option). Begin by cloning the project, installing dependencies, and starting the MCP server.

# Clone the MCP server repository
git clone https://github.com/yourusername/youtube-knowledge-base-mcp.git
cd youtube-knowledge-base-mcp

# Install the MCP runtime if needed (example using uvx tooling)
uv sync

# Prepare a working environment file (optional but recommended)
cp .env.example .env

Additional configuration and usage notes

Configure environmental data location and runtime options to fit your setup. You can place data in the OS default application data directories, or customize the path with an environment variable.

Available tools

process_video

Add a video to the knowledge base with optional tags and summary, ingesting its transcript and metadata.

manage_source

Update tags and summary for a source to refine organization and searchability.

explore_library

Browse sources, list tags, and view statistics about your library.

search

Perform a hybrid semantic plus keyword search with reranking to locate relevant moments in videos.