An AI studio
for YouTube Shorts.
Describe a Short. ytFactory writes the script, casts the voice, generates the visuals, cuts the captions, and ships it. Eight live channels, every knob you need.
no signup · single operator · runs on your GCP
Built for the work, not the demo.
Every part of a Short — script, voice, visuals, captions, publish — is a knob, a stage, a fallback. No black boxes.
Every knob the channel exposes.
Voice, length, visual style, music bed, captions density, source URL, schedule. Override only what matters; the studio handles the rest.
One pipeline, eight templates.
Rewrite → cast → images → TTS → ASR → compose → upload. Stage timeline streams to the queue.
Cloned, cached, on tap.
F5, Chatterbox, Kokoro on Cloud Run L4. ~10× faster than laptop with automatic local fallback.
FLUX.2 klein, always warm.
Per-channel style prompts on Cloud Run with min-instances=1. Render stalls fall back to local Z-Image-Turbo.
ASR-locked. Word-perfect.
Whisper-MLX timestamps every word; captions overlay at compose. No baked-in text in diffusion.
API first. Studio fallback.
When YouTube's 10K-unit quota burns through, Playwright drives Studio for unattended uploads. Same channel, no wait.
From idea to published in three clicks.
Eight battle-tested channels. Each one tuned for its niche — voice, length, visual grammar.
Override the bits you care about. Topic, source, voice, length, music, captions, schedule.
Watch the stage timeline live. Preview, approve, publish. API or Playwright fallback when quota's done.