Skip to content

PraisonAI Editor

AI-powered audio & video editor — transcribe, clean, and edit media with one command.

  • Transcribe

    Word-level timestamps in SRT, TXT, or JSON. OpenAI Whisper + local faster-whisper.

    transcribe

  • Edit

    Remove fillers, repetitions, silences, or keep only singing. One command.

    edit

  • Stem Separation

    Powered by Demucs — isolate vocals from instruments to find singing zones.

    demix

  • AI Agent

    "Remove the intro and any off-topic discussion about weather" — plain English.

    prompt edit

How it works

flowchart LR
    A[🎵 Input\nMP3 · MP4 · WAV] --> B[Probe\nmetadata]
    B --> C[Transcribe\nWhisper API]
    C --> D[Detect\nContent blocks]
    D --> E[Plan\nkeep / remove]
    E --> F[Render\nFFmpeg]
    F --> G[🎧 Output\nedited file]
    C -.->|cache| C
    D -.->|Demucs stems cache| D

Install in 30 seconds

pip install praisonai-editor
export OPENAI_API_KEY=sk-...

praisonai-editor edit podcast.mp3 -v

Feature overview

Feature Command Extra install?
Probe metadata probe No
Convert format convert No
Transcribe (Whisper) transcribe No
Create edit plan plan No
Edit (podcast/meeting/course…) edit No
Content detection (ensemble) edit --detector ensemble No
INA speech segmenter edit --detector ina [detect]
Stem separation (Demucs) edit --demix [demix]
Primary singing zone crop edit --demix --primary-zone [demix]
AI agent editing edit --prompt "…" No
Local Whisper edit --local [local]