Skip to main content

Quickstart

One API key, one recording, and something for a kitten to say. By the end of this page, you will have made both an STT and a TTS request.

Prerequisites​

1. Set your API key​

export KITTENML_API_KEY="sk_kitten_live_..."

Keep the complete key in a server-side secret. Do not put it in source code, browser JavaScript, mobile binaries, logs, or URLs.

2. Transcribe an audio file​

First, let EmoKitten listen.

Replace speech.wav with a path on your machine:

curl --fail-with-body -sS \
  https://api.kittenml.com/v1/audio/transcriptions \
  -H "Authorization: Bearer $KITTENML_API_KEY" \
  -F "file=@speech.wav" \
  -F "model=emokittenasr-realtime" \
  -F "response_format=verbose_json"

The @ tells cURL to read the file from disk. Relative paths are resolved from your current directory; absolute paths work as well.

The response includes a clean transcript, an enriched transcript, language, duration, and speech segments. See Upload transcription for every input field and response format.

3. Generate speech​

Now make a kitten talk.

curl --fail-with-body -sS \
  https://api.kittenml.com/v1/audio/speech \
  -H "Authorization: Bearer $KITTENML_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
    "model": "kitten-tts-mini-0.8",
    "input": "Hello! This audio was generated by KittenTTS.",
    "voice": "Bella",
    "response_format": "mp3",
    "stream_format": "audio",
    "speed": 1.0
  }' \
  --output kitten-output.mp3

The success body is chunked binary audio. cURL writes each chunk to the file as it arrives; do not pipe binary output to jq or print it in a terminal.

What happened?​

  • The bearer token identified your KittenML organization.
  • Upload STT processed one complete audio file.
  • KittenTTS streamed one MP3 at a price of $0.
  • Each response included an x-request-id header for tracing and support.

If both commands worked, you are in. Swap in your own audio, try another voice, and make some noise.

Continue​

Stream realtime transcriptsSend microphone audio over WebSocket and receive live results.Use an OpenAI SDKConnect familiar Python and JavaScript clients to KittenML.Clone runnable examplesTry complete integrations with sample audio and expected output.
Last updated on