Kirbah Mcp Youtube MCP server
{
"mcpServers": {
"smithery-kirbah-mcp-youtube": {
"type": "http",
"url": "https://server.smithery.ai/@kirbah/mcp-youtube/mcp"
}
}
}{
"mcpServers": {
"smithery-kirbah-mcp-youtube": {
"type": "http",
"url": "https://server.smithery.ai/@kirbah/mcp-youtube/mcp"
}
}
}{
"mcpServers": {
"smithery-kirbah-mcp-youtube": {
"type": "http",
"url": "https://server.smithery.ai/@kirbah/mcp-youtube/mcp"
}
}
}{
"servers": {
"smithery-kirbah-mcp-youtube": {
"type": "http",
"url": "https://server.smithery.ai/@kirbah/mcp-youtube/mcp"
}
}
}{
"mcpServers": {
"smithery-kirbah-mcp-youtube": {
"type": "http",
"url": "https://server.smithery.ai/@kirbah/mcp-youtube/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 0.2.6
What it does
This MCP server provides structured YouTube data access optimized for LLM token efficiency. It exposes YouTube API functionality while reducing token consumption by up to 87% compared to raw API responses, and includes optional MongoDB caching to protect daily API quotas. It is designed for AI agents and applications that need reliable YouTube data without excessive context window usage.
- Fetch video details, channel statistics, and search results with minimal token overhead for LLM processing.
- Read video transcripts directly without requiring a YouTube API key in zero-config mode.
- Cache repeated YouTube API requests to MongoDB to avoid quota exhaustion on the 10,000 daily limit.
Tools this server exposes
| Tool | Does |
|---|---|
| getVideoDetails | Retrieve structured details about a specific YouTube video. |
| getChannelStatistics | Get aggregated statistics for a YouTube channel. |
| searchVideos | Search YouTube for videos and retrieve results in a token-optimized format. |
| getTranscript | Read the transcript of a YouTube video by video ID. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.