Skip to main content

​
Overview

Voice cloning allows you to create highly realistic custom AI voices from your own audio samples. This enables a personalized and branded voice experience for your applications.
  1. Upload Audio: Provide clear audio samples of the voice you want to clone
  2. Generate Voice: Outspeed’s AI processes the samples to create a new voice
  3. Use in Apps: Integrate your custom voice into your agents and TTS requests
Cloned voices are currently only available for use with our Text-to-Speech (TTS) API.

​
Getting Started

  1. Visit Dashboard: Go to the Voice Upload page in your Outspeed Dashboard
  2. Upload Samples: Follow the instructions to upload your audio files
  3. Create Voice: Once processed, your new voice will be available to use
  4. Copy Voice ID: Copy the generated Voice ID for use in your code

​
Audio Requirements

  • Length: Minimum 10 seconds of clear speech
  • Size: Maximum 3 MB
  • Quality: High-quality audio (e.g., studio recording, no background noise)
  • Format: Only WAV files are supported
  • Content: Natural speech, varied tone and pitch
Do not upload copyrighted material or impersonate individuals without consent. Ensure you have the necessary rights for all uploaded audio.

​
Usage

Once you have your custom Voice ID, use it with our Text-to-Speech (HTTP):

​
In TTS API Requests

{
  "model": "outspeed-tts-v2",
  "voice": "your-custom-voice-id",
  "text": "This is my custom voice."
}

​
In Dialogue Generation

{
  "model": "outspeed-tts-v2",
  "text": "*Narrator:* This is my custom voice. Character: Hello!",
  "speaker_voice": "your-custom-voice-id",
  "narrator_voice": "another-custom-voice-id"
}

​
Best Practices

  • High-quality audio: Crucial for best results
  • Consent: Always obtain consent if cloning a real person’s voice
  • Monitor usage: Track how your custom voices perform in real applications