Quick Start¶

Get from zero to an edited file in under 2 minutes.

1. Install¶

pip install praisonai-editor

export OPENAI_API_KEY=sk-...

pip install "praisonai-editor[local]"
praisonai-editor edit podcast.mp3 --local

2. Edit a podcast¶

praisonai-editor edit podcast.mp3 -v

What this does:

sequenceDiagram
    participant You
    participant CLI
    participant Whisper
    participant FFmpeg

    You->>CLI: edit podcast.mp3
    CLI->>Whisper: transcribe (word-level)
    Whisper-->>CLI: words + timestamps
    CLI->>CLI: detect fillers · repetitions · silences
    CLI->>FFmpeg: render keep-only segments
    FFmpeg-->>You: podcast_edited.mp3 ✓

3. Common recipes¶

praisonai-editor edit episode.mp3 --preset podcast -v

praisonai-editor edit concert.mp3 \
  --preset songs_only \
  --detector ensemble \
  --demix \
  --primary-zone \
  -v

praisonai-editor edit standup.mp4 --preset meeting -v

praisonai-editor edit interview.mp3 \
  --prompt "Remove the intro, the weather tangent, and keep only the technical discussion"

praisonai-editor transcribe video.mp4 --format srt --output subs.srt

praisonai-editor convert video.mp4 --format mp3

4. Output files¶

After edit, you get:

podcast_edited.mp3        ← edited audio
~/.praisonai/editor/podcast/
  ├── probe.json          ← media metadata
  ├── transcript.json     ← word timestamps (cached)
  ├── transcript.srt      ← subtitles
  ├── transcript.txt      ← plain text
  ├── plan.json           ← edit decisions
  └── content_blocks.json ← detection blocks (if --detector)