Start
Quickstart
Create a key, check your balance, and send your first request in a few minutes.
- 1
Create an API key
Sign in, open API keys, choose New Key, name it, and choose Create Key. Copy the secret straight away. It starts with
sk-quan-and is shown only once. - 2
Check your balance
Kael is prepaid. A request is refused with a
402error when your balance is $0 or below. Your balance is on the dashboard. To add credit, see Pricing and billing. - 3
Send a request
Use the OpenAI SDK you already have, or plain HTTP. Only the base URL and the model name change. Kael also speaks the Anthropic Messages format; the last tab shows it.
curl https://api.quancis.space/v1/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $QUAN_API_KEY" \ -d '{ "model": "kael-beta", "messages": [ {"role": "user", "content": "What is wrong with this loop? for i in range(len(xs) + 1): total += xs[i]"} ] }'
Read the response
The answer is in choices[0].message.content. The usage object shows exactly what you were billed for: prompt_tokens in, completion_tokens out, and how many input tokens were served from cache.
{
"id": "...",
"object": "chat.completion",
"created": 1790000000,
"model": "kael-beta",
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "The range goes one step too far: ..."
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 38,
"completion_tokens": 214,
"total_tokens": 252,
"prompt_cache_hit_tokens": 0
}
}Next
- Use a different format or route: Endpoints.
- Make Kael think harder on a hard request: Thinking levels.
- Handle failures: Errors.