Glossary
VOD Ingest
VOD ingest is the pipeline stage that fetches a video-on-demand source — a Twitch VOD, YouTube upload, Kick replay — and prepares it for clipping.
Ingest is the unglamorous foundation of every clipping product. Before any AI runs, the source has to be downloaded, transcoded to a known codec and frame rate, transcribed, and optionally cookie-bypassed past platform anti-scraping checks. Production pipelines normalize to H.264, 30 fps, 720p or 1080p depending on tier, since downstream face detection and reframing assume a fixed input shape.
Ingest is also where most pipelines fail. YouTube rotates anti-bot signatures every few weeks, Twitch enforces VOD download restrictions for sub-only content, and Kick's API rate-limits aggressively. A robust ingest layer needs cookie pools, multi-IP egress, and per-platform fallback chains to keep success rates above 95%.
AutoClip's VOD ingest worker runs on Modal with a 3-path fallback: Tailscale-routed yt-dlp first, curl_cffi with Chrome impersonation second, cookie-authenticated direct download as the last resort. Average ingest time for a 60-minute YouTube source is 90-180 seconds.
Related Terms
Frequently Asked Questions
What VOD sources does AutoClip support?
YouTube (including unlisted), Twitch VODs, Kick replays, and direct MP4/MOV uploads. Twitch sub-only VODs require the clipper to provide their own session cookie.
How long does VOD ingest take?
Roughly 1.5-3 minutes for a 60-minute source on a fast egress path. Slow paths (cookie fallback, throttled connections) can stretch to 5-8 minutes.
Put VOD Ingest to Work
AutoClip handles the full pipeline — viral moment detection, 9:16 reframing, captions, and auto-posting. Start clipping for free.
Get Started Free