Finding footage. Turning speech into text.
Building a place for research to accumulate.
A workspace for finding, downloading and choosing footage for an edit.
The search for one useful shot can leave dozens of tabs, scattered downloads and decisions that are hard to revisit.
Python / Local web interface
Local audio and video transcription, with a batch queue and optional speaker separation.
Recorded conversations and long videos are difficult to search, quote and reuse while their content stays inside the media file.
WhisperX / FastAPI / NVIDIA CUDA
A planned research workspace that builds a reusable foundation for long-form video production.
Research loses value when sources, transcripts and decisions are scattered across chats and gathered again for every new script.
MCP · planned integration