Quickstart
Make your first speech request against the local VoiceStudio API.
Local API only
This targets the desktop app's 0.5.0 API. The Cloud API is not released and
works differently: see the Cloud preview.
1. Start the app
Open VoiceStudio. Its API listens on port 3900 while the app runs.
export VOICESTUDIO_BASE_URL="http://localhost:3900/v1"
export VOICESTUDIO_API_KEY="local"
curl http://localhost:3900/health{
"status": "ok",
"device": "cuda (NVIDIA GeForce RTX 4090)",
"version": "0.5.0"
}On localhost the key can be any non-empty string. It is only checked if you turn on remote access.
2. Generate speech
curl "$VOICESTUDIO_BASE_URL/audio/speech" \
-H "Authorization: Bearer $VOICESTUDIO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "tts-1",
"voice": "alloy",
"input": "VoiceStudio is ready.",
"response_format": "mp3"
}' \
--output speech.mp33. Or use an OpenAI SDK
import os
from openai import OpenAI
client = OpenAI(
base_url=os.environ["VOICESTUDIO_BASE_URL"],
api_key=os.environ["VOICESTUDIO_API_KEY"],
)
audio = client.audio.speech.create(
model="tts-1",
voice="alloy",
input="The local VoiceStudio API is ready.",
)
audio.stream_to_file("speech.mp3")Next: swap alloy for one of your own voice profile IDs.