Turn YouTube, TikTok, Vimeo into structured training data. Direct S3 delivery, metadata extraction, and audio processing at scale.
Leverage the power of yt-dlp/ffmpeg without managing infra. We handle proxies, updates, and scaling.
Get rich JSON metadata including title, description, views, likes, comments, and subtitles.
We push files directly to your S3/GCS bucket. Save bandwidth and simplify your pipeline.
PB-scale downloader clusters capable of handling millions of video jobs concurrently.
# Example: Batch Download to S3
curl -X POST "https://api.123proxy.com/v1/video/extract" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
"output_format": "mp4",
"s3_config": {
"bucket": "my-training-data",
"region": "us-east-1",
"access_key": "YOUR_KEY",
"secret_key": "YOUR_SECRET"
},
"fetch_metadata": true
}'
The API asynchronously processes the video, handles all network issues, and uploads the final clean file and metadata JSON to your bucket. You receive a webhook upon completion.
Built-in cleaning filters for noise free data.
Pay-as-you-go per successful request.
Comprehensive documentation, SDKs for Python/Node.js, and webhooks for
integration.
Industry leading 99.2% success rate on difficult targets.
Only pay for valid 200 OK responses.
Built for high-scale AI data ingestion.
| Feature | Specification |
|---|---|
| Supported Platforms | YouTube, TikTok, Vimeo, Instagram, Twitter, Bilibili + 1000 others |
| Output Formats | MP4, WEBM, MKV, MP3, WAV, M4A, JSON (Metadata) |
| Delivery Methods | Direct S3 Upload (AWS), GCS, Webhook Callback |
| Metadata Fields | Title, Description, Views, Likes, User Info, Comments, Subtitles (VTT/SRT) |
| Concurrency | Unlimited concurrent async jobs (auto-scaling) |
| Proxy Network | 72M+ Residential IPs, ISP Proxies for 99.9% Success Rate |
Yes. You can submit a playlist URL or a shorts URL. For playlists, we can expand it and schedule jobs for each video automatically.
You can specify `output_format: "mp3"` or `"wav"` in your request. We will extract the audio track and convert it to your desired format before uploading to S3.
We support any platform supported by yt-dlp, which includes YouTube, TikTok, Vimeo, Bilibili, Instagram, Twitter, and over a thousand others.
Yes. You can set `metadata_only: true` to fetch only the JSON metadata details like views, likes, description, tags, and subtitles without processing the media file.
We offer high concurrency limits. Since jobs are asynchronous, you can submit thousands of jobs at once, and we process them in parallel using our massive proxy network.