Agents Lookup

Use cases / Decision models (Jev)

Label short-form videos by hook, format and angle for content research

Turn each video in a niche into a transcript, have Jev label the opening, structure and angle, then search the library by structure.

DifficultyIntermediate
Setup time3-6 hours
CategoryDecision models (Jev)

Author's figure: Author's figure: 1,066 reaction videos analysed with Jev and Gemini (Meet Shukla). Reported by the builder, not measured by us.

The use case

Content teams studying a niche, and agencies building swipe libraries, cannot watch thousands of short videos. Each video's transcript and caption, sometimes with a frame description from a vision model, goes to Jev with Choice questions on opening type, structure, angle and presenter. The library becomes searchable by structure, so a team can pull every reaction clip that opens with a question. A writer or strategist watches the examples before scripting anything, and reuse of other people's clips must respect copyright and platform rules.

Example workflow

  1. Collect videos and captions with a TikTok or YouTube data source
  2. Transcribe audio with Whisper and describe key frames with a vision model
  3. Ask Jev Choice questions on opening, structure, angle and presenter
  4. Store labels in Airtable for search
  5. A strategist watches examples before writing scripts

Tools

Risks and guardrails

Sources

Written by Agents Lookup. Credits and official pages for the tools this use case names: