youtube-mcp
by kzmshx
README.md
# kzmshx-youtube-mcp
An MCP server for extracting YouTube video transcripts and metadata.
## Configuration
```json
{
"mcpServers": {
"youtube": {
"command": "uvx",
"args": ["kzmshx-youtube-mcp"]
}
}
}
```
## Installation (Optional)
If you prefer to install globally:
```bash
pip install kzmshx-youtube-mcp
# or
uv tool install kzmshx-youtube-mcp
```
## Tools
### get_transcript
Get transcript text from a YouTube video.
| Parameter | Type | Description |
| ----------------- | ------ | ---------------------------------------- |
| `url` | string | YouTube URL or video ID |
| `language` | string | Language code (default: auto) |
| `with_timestamps` | bool | Include timestamps (default: false) |
**Example:**
```json
// Input
{ "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ" }
// Output
{
"video_id": "dQw4w9WgXcQ",
"language": "en",
"transcript": "We're no strangers to love\nYou know the rules and so do I..."
}
```
### get_video_info
Get metadata from a YouTube video.
| Parameter | Type | Description |
| --------- | ------ | ----------------------- |
| `url` | string | YouTube URL or video ID |
**Example:**
```json
// Input
{ "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ" }
// Output
{
"id": "dQw4w9WgXcQ",
"title": "Rick Astley - Never Gonna Give You Up",
"description": "...",
"duration": 213,
"channel": "Rick Astley",
"upload_date": "2009-10-25",
"view_count": 1500000000,
"available_languages": ["en", "ja", "es", "fr"]
}
```
## License
MIT
TDQS
B3.2/5.0
Scored across 2 tools
Disambiguation5/5
The two tools target clearly different outputs: transcript text versus video metadata. There is no realistic chance of an agent confusing them.
Naming Consistency5/5
Both tools use the same get_<noun> snake_case pattern, which is perfectly consistent.
Tool Count3/5
Two tools is thin for a server, though each one earns its place for a transcript-and-metadata use case. It feels more like a minimal utility than a full YouTube surface.
Completeness3/5
Core read operations for a single video are covered, but there is no way to search videos, list channel or playlist content, or discover videos, so agents cannot go from a topic to a video ID. For a read-only extraction tool this is workable but has notable gaps.
Maintenance
ActivityNo data
ResponsivenessOutdated data