v2.0.0 — open source

VideoCrab

Drop a 2-hour video. Get a detailed breakdown of everything discussed — with timestamps, topics, and clips. Powered by AI.

Requires FFmpeg · Free & open source · MIT License

What it does

Drop a video.
Get everything out of it.

01 Transcribe Whisper-powered speech-to-text. Segment by region, navigate by timestamp, multi-language.

02 Analyze 3-stage context-aware AI pipeline. Detailed narrative, topics, key moments. Ask about specific aspects with voice.

03 Trim & Clip Waveform timeline, drag-to-select, AI-suggested clip boundaries. Export with FFmpeg.

04 Listen TTS narration of analyses. Section-based streaming, built-in player, per-session caching.

05 Organize Projects with auto-save. All transcripts and analyses persist. Pick up where you left off.

See it in action

How it works

The 3-stage analysis pipeline ensures accurate understanding of your content, even for multi-hour recordings.

AI reads the full transcript to understand participants, topic, and narrative arc.

10-minute chunks analyzed in parallel, each with full video context for accuracy.

Chunk analyses merged into one cohesive narrative with topics, moments, and clips.

Free, open source, runs locally on your machine.