Temp File Output: Reduced Context Usage

⚠️ BEHAVIOR CHANGE ⚠️

This release changes how transcript data is returned to reduce context usage.

Temp File Output

Full transcript JSON is now saved to a temp file instead of returned inline.

What changes:

• The get_transcript tool now returns a summary with metadata and preview
• Full transcript saved to /tmp/youtube-transcript-{video_id}.json
• Significantly reduced context window consumption for long videos

Response format:

Successfully extracted transcript for video abc123
Language: English (en)
Generated: True
Segments: 150
Characters: 25,000

Preview (first 3 segments):
1. [0.0s] Hello and welcome...
2. [3.5s] Today we'll be...
3. [7.2s] Let's get started...
... (147 more segments)

Full transcript saved to: /tmp/youtube-transcript-abc123.json

Installation

pip install youtube-transcript-mcp==0.3.0

Rollback

pip install youtube-transcript-mcp==0.2.0

See RELEASE_NOTES_v0.3.0.md for full details.