Footage intelligence search
semantic search across every frame and word
Search terabytes of footage by describing what you need in plain English. EditAssist indexes every frame with local vision models and every word with Whisper-based transcription. Results are ranked by relevance and ready to drop into the timeline.
Workflow
Ingest your media library and the agent indexes every frame with local vision embeddings and transcribes every word with Whisper. You can then search by visual description, transcript phrase, speaker, camera metadata, or emotional tone. Results appear in seconds across terabytes of footage.
Results are ranked by semantic relevance and narrative fit. The agent can build a selects timeline from search results, suggest cutaway placements, or cross-reference footage against scripts and paper edits. All indexing is local and private.
Prompts
“Find every close-up of the lead actor across all three camera angles.”
“Pull all B-roll of the city skyline and arrange by time of day.”
“Find the three best takes of the monologue and lay them side by side for the director.”
“Find every clip shot at golden hour and arrange them chronologically.”
“Search for every clip shot on the Alexa Mini and filter by ISO above 1600.”
“Find every exterior shot and check if any are overexposed by more than a stop.”
“Find all clips where the boom is visible in frame and flag it for the editor.”
“Cross-reference 200 hours of archival footage against the paper edit and pull matching clips.”
Features
Visual & transcript indexing
Every frame is embedded with local vision models; every word is transcribed with Whisper. Search across both simultaneously.
Semantic ranking
Results ranked by visual relevance, narrative fit, audio energy, and transcript punchiness, not just keyword matching.
Auto-selects timeline
Builds a dedicated selects timeline from search results, ready to drop into the main edit as cutaways or inserts.
Local & private
All indexing runs on your own GPU. No footage uploaded. No cloud dependency. No per-minute fees.
Local-first
All transcription, vision analysis, semantic search, and clustering run on your own GPU. Nothing about your footage is uploaded. The only data that leaves is the text of your conversation with the agent. It's free to download, local models are free, and a new account starts with £15 of credit and no card. For long-running cloud work, pay as you go, or choose £20/mo Starter or £100/mo Pro.
Getting started
- 01Download EditAssist for macOS or Windows and create an account. You get £15 credit and no card is needed.
- 02Ingest your footage. The agent transcribes, indexes, and clusters automatically.
- 03Describe what you want in plain English and let the agent build the cut.
- 04Review, refine with conversational commands, and export to your delivery spec.
More tools: AI B-roll finder, AI transcript editor, Rough cut generator. Or see everything EditAssist connects to.