Repurposing 60-minute technical keynotes and podcasts into viral 9:16 shorts is painful. Generic stock footage from Pexels is irrelevant, expensive, and destroys retention
We built Dissect
Dissect is an autonomous video re-cutter that ingests long-form keynotes and generates context-aware cinematic 9:16 B-roll using Livepeer's decentralized GPU network.
How it works:
- Word-Level Transcription & Retention Spikes
Dissect transcribes speech using Livepeer MCP, mapping audience retention curves and flagging talking-head dead zones (>3s) where viewers drop off
- Contextual 9:16 B-Roll Synthesis
Instead of pulling unrelated stock footage, Dissect translates spoken arguments into vertical camera directives and dispatches parallel diffusion jobs across decentralized Livepeer nodes
- Interactive 3-Track NLE
Reframed speaker footage, generative B-roll cutaways, and kinetic captions play in locked 60 FPS inside an interactive smartphone simulator with real-time facial reticles
- Conversational Slice Surgery
Scrub to any cut on the timeline. Click to swap camera directives (macro, prime, drone) and re-render the slice on Livepeer in seconds without touching the rest of the cut
- Direct Multi-Platform Export
Audit real-time GPU compute telemetry and export finalized reels with adaptive safe zones for TikTok, Shorts, and Reels
Built with @Livepeer for @Embody_Net Hackathon
check it out here: https://t.co/Zh9oX4YvJW
$10 $SOL | 1luckywinner |xpicker.
- join: https://t.co/OrZ899DuAS
- show proof & drop address
go to channel & claim your spot now.
ALL GIVEAWAYS WILL NOW CONTINUE IN DISCORD.
24 ⏰