Problem
Documentary and editorial projects spread footage across drives, making it slow to find quotes, understand the story across hours of material, and prepare an edit without duplicating large media files.
Organize source media by reference, search transcripts and analysis, and build an editable project guide with quotes, emotional and narrative maps, subtitles, timelines, and selects.

Project anatomy
Documentary and editorial projects spread footage across drives, making it slow to find quotes, understand the story across hours of material, and prepare an edit without duplicating large media files.
PantoScan indexes media in place, transcribes it locally with word-level timestamps, and uses a local LLM to map topics, emotions, and narrative roles. Its searchable, editable guide exports as Markdown, SRT, FCPXML, and selects CSV.
Main features
Index external media by volume and relative path, without moving or duplicating files
Batch processing with job recovery, cancellation, and reuse when source fingerprints are unchanged
Local whisper.cpp transcription with word timestamps and local LLM analysis of 11 topics, 11 emotions, and narrative roles
Search clips, topics, emotions, and quotes across the library; edit results and regenerate outputs
Export project guides as Markdown, subtitles as SRT, timelines as FCPXML, and selects as CSV
Electron
macOS
whisper.cpp
FFmpeg
Local LLM
SQLite FTS5
SRT
FCPXML
Project media
