DarkString / Products
VogCut: Local Video Search and AI Video Understanding
VogCut is for video teams that need to find useful material in an existing library. It indexes local footage, analyzes scenes and dialogue, and organizes the results into searchable clips and editable cut plans.
Find the content before planning the edit
File names and folders cannot always describe what happens in a shot. VogCut focuses on understanding and retrieving existing footage. For planning newly generated shots, explore Tapdance.
Move from a library to useful clips
Start with representative footage and check whether indexing, retrieval and cut planning support the tasks your team performs.
- Define the local directories, formats and footage to process.
- Analyze scenes and dialogue to support clip retrieval.
- Organize retrieved material into an editable cut plan.
Validate search quality with real footage
Image quality, language, audio and shot types affect understanding. Prepare typical queries and known target clips. Local storage does not imply every inference step is offline: model calls, hardware requirements and data flows depend on the current configuration.
Questions and answers
Does it replace every part of manual editing?
The focus is footage understanding, retrieval and cut planning. Narrative, pacing and final output still require project-specific review.
What is needed for an evaluation?
Bring representative footage, common search questions, expected outputs and device constraints. Confirm permissions and data boundaries before using unpublished material.
Explore further
Share the audience, available assets, device or model constraints, and the task you want to complete.
Discuss a project →