Overview
QuickVoice is two endpoints: one that turns text into speech, one that turns speech into text. Both stream, both take the same bearer key, and both bill per character to four decimal places.
Base URL
Every request goes to a single host. There are no regional subdomains to pick between.
https://api.quickdial.ai
Authentication
Authenticate with a bearer token in the Authorization header. The same key works for
text-to-speech, speech-to-text and both WebSocket stream endpoints. You do not need separate
credentials per service.
Authorization: Bearer qtts_live_xxxxxxxxxxxxxxxxxxxx
Create and rotate keys from the dashboard. Keys are secret, so treat them like a password and keep them server-side. A browser that calls the API directly will leak the key to anyone who opens devtools; proxy through your own backend instead.
Your first request
This returns an Opus audio stream and writes it to a file:
curl https://api.quickdial.ai/v1/tts \ -H "Authorization: Bearer $QUICKVOICE_KEY" \ -H "Content-Type: application/json" \ -d '{"text":"I found three options that match your request.","voice":"azelma"}' \ --output speech.opus
How billing works
- Text-to-speech: $0.0028 per 1,000 characters of input text (Opus). WAV output is $0.0045 per 1,000 characters, reflecting the larger lossless payload.
- Speech-to-text: $0.0018 per 1,000 characters of transcribed output.
- Usage is metered to four decimal places and shown per request in the dashboard, so an invoice reconciles to the cent.
- New accounts get 100,000 characters free, roughly 100 minutes of speech. No card required, and no monthly minimum after that.
Concurrency and uptime
Workers are stateless and designed to scale horizontally. Request rate is governed per key and reported in response headers, so separate workloads can be isolated cleanly. See Errors & limits for the headers that report the current request budget.
Next
- Quickstart: a working script in each language.
- Text to Speech: parameters, voices and formats.
- Speech to Text: uploads, timestamps and response shape.
- Streaming: the WebSocket protocol for both directions.