Runnable examples

The KittenML API examples repository contains small programs and sample audio for every primary inference route.

Clone them. Run them. Change them. Break them. These are deliberately small enough to understand before you carry them into a real application.

Task	cURL	Python	Node.js
Complete-response TTS	Yes	OpenAI SDK	OpenAI SDK
Binary or SSE output-streaming TTS	Yes	OpenAI SDK	OpenAI SDK
Realtime text-input TTS	—	WebSocket	WebSocket
List TTS voices	Yes	Standard library	fetch
Kitten TTS 2 decoding controls	Yes	OpenAI SDK	OpenAI SDK
Upload STT	Yes	OpenAI SDK	OpenAI SDK
Upload STT with SSE	Yes	—	—
Realtime STT	—	websockets	ws
Multi-turn realtime STT	—	websockets	—
Browser WebRTC STT	—	—	Browser + Node.js token server

Clone and configure

git clone https://github.com/KittenML/kittenml-api-examples.git
cd kittenml-api-examples
export KITTENML_API_KEY="sk_kitten_live_..."

The repository includes a small audio sample, dependency versions, expected terminal output, and a README beside each example.

Pick your path

• TTS examples
  • Complete speech response
  • Binary and SSE output streaming
  • Realtime text input
  • Voice listing
  • Decoding controls
• STT examples
  • Upload transcription
  • Realtime WebSocket clients
  • Browser WebRTC client
• Realtime event reference

Happy building, kittens. If an example behaves differently on your audio or in your environment, keep the request_id and tell us what happened.