athina-ai/youtube-apify-transcript

youtube-apify-transcript

Fetch YouTube transcripts via APIFY API (works from cloud IPs, bypasses YouTube bot detection).

Why APIFY?

YouTube blocks transcript requests from cloud IPs (AWS, GCP, etc.). APIFY runs the request through residential proxies, bypassing bot detection reliably.

Free Tier

• $5/month free credits (~714 videos)
• No credit card required
• Perfect for personal use

Cost

• $0.007 per video (less than 1 cent!)
• Track usage at: https://console.apify.com/billing

Links

• APIFY Pricing
• Get API Key
• YouTube Transcript Scraper Actor

Setup

1. Create free APIFY account: https://apify.com/
2. Get your API token: https://console.apify.com/account/integrations
3. Set environment variable:

# Add to ~/.bashrc or ~/.zshrc
export APIFY_API_TOKEN="apify_api_YOUR_TOKEN_HERE"

# Or use .env file (never commit this!)
echo 'APIFY_API_TOKEN=apify_api_YOUR_TOKEN_HERE' >> .env

Usage

Basic Usage

# Get transcript as text (uses cache by default)
python3 scripts/fetch_transcript.py "https://www.youtube.com/watch?v=VIDEO_ID"

# Short URL also works
python3 scripts/fetch_transcript.py "https://youtu.be/VIDEO_ID"

Options

# Output to file
python3 scripts/fetch_transcript.py "URL" --output transcript.txt

# JSON format (includes timestamps)
python3 scripts/fetch_transcript.py "URL" --json

# Both: JSON to file
python3 scripts/fetch_transcript.py "URL" --json --output transcript.json

# Specify language preference
python3 scripts/fetch_transcript.py "URL" --lang de

Caching (saves money!)

Transcripts are cached locally by default. Repeat requests for the same video cost $0.

# First request: fetches from APIFY ($0.007)
python3 scripts/fetch_transcript.py "URL"

# Second request: uses cache (FREE!)
python3 scripts/fetch_transcript.py "URL"
# Output: [cached] Transcript for: VIDEO_ID

# Bypass cache (force fresh fetch)
python3 scripts/fetch_transcript.py "URL" --no-cache

# View cache stats
python3 scripts/fetch_transcript.py --cache-stats

# Clear all cached transcripts
python3 scripts/fetch_transcript.py --clear-cache

Cache location: .cache/ in skill directory (override with YT_TRANSCRIPT_CACHE_DIR env var)

Batch Mode

Process multiple videos at once:

# Create a file with URLs (one per line)
cat > urls.txt << EOF
https://youtube.com/watch?v=VIDEO1
https://youtu.be/VIDEO2
https://youtube.com/watch?v=VIDEO3
EOF

# Process all URLs
python3 scripts/fetch_transcript.py --batch urls.txt

# Batch with JSON output to file
python3 scripts/fetch_transcript.py --batch urls.txt --json --output all_transcripts.json

APIFY Actor Input

The script sends the following input to pintostudio/youtube-transcript-scraper:

{
  "videoUrl": "https://www.youtube.com/watch?v=VIDEO_ID"
}

Output fields:

Each result contains a data array of transcript segments:

| Field | Type | Description | |---------|--------|------------------------------------| | start | number | Segment start time (seconds) | | dur | number | Segment duration (seconds) | | text | string | Transcript text for this segment |

Output Formats

Text (default):

Hello and welcome to this video.
Today we're going to talk about...

JSON (--json):

{
  "video_id": "dQw4w9WgXcQ",
  "title": "Video Title",
  "transcript": [
    {"start": 0.0, "dur": 2.5, "text": "Hello and welcome"},
    {"start": 2.5, "dur": 3.0, "text": "to this video"}
  ],
  "full_text": "Hello and welcome to this video..."
}

Error Handling

The script handles common errors:

• Invalid YouTube URL
• Video has no transcript
• API quota exceeded
• Network errors