Docs/AngeVoice

Start using AngeVoice

This page takes you from a prepared host to a WAV file you can play and verify.

DockerFirst synthesis10–20 minutes

Before you start

Choose a deployment

HostRecommended profileDefault port
No NVIDIA GPU, NAS or first validationcpu8100
Regular NVIDIA x86_64 hostgpu8101
Older driver environment where the standard GPU image failslegacy-gpu8102

Try the standard GPU profile first. Legacy GPU is a compatibility option, not a faster profile.

Run the installer

bash <(curl -fsSL https://raw.githubusercontent.com/AngeVox/AngeVoice/main/scripts/install.sh)

If GitHub access is slow, clone the repository first:

git clone https://github.com/AngeVox/AngeVoice.git
cd AngeVoice
bash scripts/install.sh

First sign-in and password change

  1. Open the Studio URL printed by the installer.
  2. Open the admin console. The first sign-in can use admin / admin123.
  3. Change the administrator name and password immediately.
  4. Open the security page, copy or rotate the API key and store it in the client securely.

The default credentials exist only to provide an initial entry path. Never keep them for a network-accessible deployment.

Verify the service

curl http://127.0.0.1:8100/health
curl http://127.0.0.1:8100/v1/models

HTTP 200 with ok or idle means the service is available. loading means the model is still starting or downloading.

Generate the first audio file

curl -X POST http://127.0.0.1:8100/v1/audio/speech \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -d '{"model":"kokoro","input":"Hello from AngeVoice.","voice":"zm_010","response_format":"wav"}' \
  --output hello.wav

Play hello.wav. A 401 response usually means the Bearer token is missing or incorrect.

Next steps