Turn one sentence into a narrated video.
Zero cloud APIs. Zero subscriptions. Your GPU does the work.
phi3.5 turns your logline into a poetic 60-second script with narration and scene descriptions.
Automatically breaks the script into shots with detailed image prompts for Stable Diffusion.
Generates cinematic images locally on your RTX 4070 (or better). No cloud image APIs.
Piper neural TTS reads your narration with a natural voice. No ElevenLabs subscription.
FFmpeg stitches images and audio into a polished MP4. TikTok vertical format included.
Generate a week's worth of content in one command. The swarm works while you sleep.
Full source code. No subscriptions. No cloud bills.
Built by a disabled founder on a $1,200 laptop.