home / mcp / youtube138 mcp server

Youtube138 MCP Server

Provides MCP access to Youtube138 API endpoints via stdio transport with API key authentication.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "bach-ai-tools-bachai-youtube138": {
      "command": "uvx",
      "args": [
        "--from",
        "bach-youtube138",
        "bach_youtube138"
      ],
      "env": {
        "API_KEY": "YOUR_API_KEY_HERE"
      }
    }
  }
}

You can access the Youtube138 API through this MCP server using a stdio transport. It provides a set of YouTube-related endpoints you can query from your application with a consistent MCP experience and centralized API key management.

How to use

Connect to the Youtube138 MCP server from your MCP client using the stdio flow. You will run the MCP server as a local process and point your client to its standard input/output interface. Before you start, set your API key so requests are authenticated.

Two common startup patterns are supported: you can use the uvx launcher to start the server from the package, or run the server directly in development mode. Ensure you provide your API key via an environment variable and then invoke the start command from your MCP client.

How to install

# Prerequisites: Python and a package manager for your environment

# Install the server package from PyPI
pip install bach-youtube138

# Start in standard runtime (example 1) using uvx if available
uvx --from bach-youtube138 bach_youtube138

# Start in standard runtime (example 2) using uvx with latest version
uvx --from bach-youtube138@latest bach_youtube138

# Or run directly in development mode
python server.py

# After installation, you can also run the package as a command
# (the command name uses an underscore)
bach_youtube138

Additional steps and configuration

Authentication requires an API key. Set the API key in your environment before starting the server or your client will be unable to authenticate requests.

If you are using MCP clients like Cursor or Claude Desktop, you can configure the server to run with the appropriate environment variables injected. Below are example configurations you can adapt to your environment.

Available tools

auto_complete

Auto Complete endpoint to fetch suggested queries for a given input.

home

Home endpoint to fetch initial content with optional cursor and region parameters.

search

Search endpoint to query content by a text string.

channel_channels

Channel channels endpoint to list channels or collections for a given channel id.

channel_search

Channel search endpoint to find channels by query.

channel_playlists

Channel playlists endpoint to retrieve playlists for a channel with filtering options.

channel_details

Channel details endpoint to fetch basic information about a channel.

channel_community

Channel community endpoint to access community posts and related data.

channel_videos

Channel videos endpoint to list videos from a channel with filters.

channel_details_v2

Channel details v2 endpoint providing extended channel metadata.

video_comments

Video comments endpoint to retrieve comments for a given video.

video_related_contents

Video related contents endpoint to fetch related videos or content.

video_streaming_data

Video streaming data endpoint to obtain streaming-related information for a video.

video_details

Video details endpoint to fetch details for a specific video.

video_details_v2

Video details v2 endpoint with extended video metadata.

community_post_comments

Community post comments endpoint to access comments on a community post.

community_post_details

Community post details endpoint to retrieve data for a community post.

playlist_videos

Playlist videos endpoint to list videos within a playlist.

playlist_details

Playlist details endpoint to fetch information about a playlist.