What problem does it solve?
Turning hours of livestream footage into publishable short clips requires watching the entire recording, identifying highlight moments, and editing each one without distorting the speaker's meaning. This Skill automates that workflow by combining transcript, visual, audio, interaction, and metadata evidence to find, score, and edit clips inside OpenChatCut.
Core Features & Use Cases
- Genre-aware section classification: Splits a recording into sections and assigns profiles such as commerce, gaming, talk, interview, education, music, sports, or IRL, each with its own event arc and packaging rules.
- Multimodal candidate discovery: Generates clip candidates from speech, visual, audio, chat interaction, and metadata signals, then applies hard rejection gates and reproducible scoring before selection.
- End-to-end editing and export: Creates each approved clip as its own timeline Sequence, applies reframing, captions, and cleanup, verifies the composed result frame by frame, and queues renders into the media pool.
- Use Case: Import a three-hour gaming livestream and ask for five vertical highlight clips; the Skill maps the stream, finds decisive plays with reactions, crops for vertical format, adds captions, verifies each cut, and exports the finished clips.
Quick Start
Import a livestream recording into the project and ask the agent to cut it into platform-ready highlight clips for your target platform.
