Quickstart

Spero is an English text-to-speech model with two US voices. You send text to one endpoint and get an audio file back.

1. Create an account

Sign up and confirm your e-mail. Your account then has $2 of credit, which covers 100,000 characters.

2. Create a key

Go to API keys and create one. It starts with sp_live_ and is shown once. Keep it on your server, never in a browser.

3. Make a request

The endpoint takes the same request as OpenAI's speech endpoint, so its client libraries work if you change the base URL.

curl https://api.spero.patientdesk.ai/v1/audio/speech \
  -H "Authorization: Bearer $SPERO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "spero-v1", "input": "Your order ships on Tuesday.", "voice": "spero-f1"}' \
  --max-time 300 \
  --output speech.mp3

How long it takes

A two-sentence passage takes about 30 seconds, and a full 2,000 characters about 3 minutes, because each request renders six takes and keeps a clean one. The audio arrives as one file, not a stream, so stream: true is refused for now. Set your HTTP client's timeout to at least five minutes. A faster engine is what we work on next.

Spero suits prerecorded audio. It does not suit live conversation today.

Try it first

The playground sends the same request from your browser and spends the same balance.

Open source

Open-source, coming soon.