Quick Start

Re-voice your first demo video in under 5 minutes.

1. Set Your API Key

DemoVoice uses OpenAI for transcription and text-to-speech. Set your API key:

export OPENAI_API_KEY=sk-your-key-here

2. Initialize the Project

Create the configuration files in your project directory:

demovoice init

This creates a .demovoice/ directory containingdemovoice.yaml and a starterglossary.yaml.

3. Inspect Your Video

Preview what DemoVoice will detect in your video:

demovoice inspect demo.mp4

This prints the detected narration segments with their timestamps and risk scores. Use it to verify the transcription before rendering — if Whisper got a term wrong, add an entry to glossary.yaml or export an editable script:

demovoice inspect demo.mp4 --output-script .demovoice/script.yaml

4. Render with the New Voice

Generate your re-voiced video. --output is required.

demovoice render demo.mp4 --output demo.demovoice.mp4

DemoVoice will transcribe the audio, synthesize new speech, fit each segment into the original timing window, and mux the new audio onto the original video stream.

5. Preview While Tuning

Full renders take time. While tuning voice, glossary, or timing settings, render a short preview:

demovoice render demo.mp4 --preview-duration 20s --output preview.mp4

6. Customize the Voice

Edit .demovoice/demovoice.yaml to change the voice or delivery:

providers:
  tts:
    provider: openai
    model: gpt-4o-mini-tts
    voice: cedar

profiles:
  default:
    voice_instructions: >
      Sound like a calm, confident technical founder.
      Keep energy moderate and delivery clear.
      Do not sound salesy.