Kaerv 0.45Support

Models API

Call Kaer models from your own code with any OpenAI-compatible SDK, create a Models API key, list models, stream replies and understand rate limits and billing.

Your Kaer account comes with an OpenAI-compatible API. Point any OpenAI SDK at it, send a key with the Models API permission, and leave the model on auto — Kaer picks the right model for each request.

PermissionModels API
Default modelauto
BillingYour credits, at each model's rate
LimitsFollow your plan
Base URL
https://app.kaer.ai/api/v1

Before you begin

  • A Kaer account with credits. Calls draw on the same credits as the app.
  • Any OpenAI SDK, or a tool that can send HTTP requests.

Create a key

Open Developers

Go to Settings → Developers.

Create a Models API key

Under Call models from your code, create a key with the Models API permission.

Copy it

Copy the key and a ready example. The key is shown once, so store it like a password.

Call a model

curl https://app.kaer.ai/api/v1/chat/completions \
-H "Authorization: Bearer kaer_sk_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "auto", "messages": [{"role": "user", "content": "Hello"}]}'
POST/api/v1/chat/completionsscope Models APIcredits
modelstring
auto (the default) or any id from GET /api/v1/models.
messagesarrayrequired
The conversation, OpenAI style: { role, content }.
streamboolean
true streams chat.completion.chunk events, ending with [DONE].

List the models you can call

GET/api/v1/models

Every model this account can call, with its price. Use any id from this list as model, or leave it on auto.

Limits

Requests a minute follow your plan — Free 10, Starter 30, Pro 60, Max 120 — and one key carries up to 60 a minute. The free plan also has a monthly allowance. Calls stop when your credits run out; top up in Settings → Billing.

Pay as you go

For company spend or heavy traffic, use api.kaer.ai. It is its own account and wallet, pay as you go, and its limits grow as your usage and top-ups do — up to 10,000 requests a minute.

Common questions

Which model does auto use?

Kaer picks the model that fits each request. To choose yourself, pass an id from GET /api/v1/models.

Can I stream the reply?

Yes. Set stream to true and read chat.completion.chunk events until [DONE].

Where do I see what my calls cost?

Settings → Usage shows spend by model and by surface. See Billing and credits.

Troubleshooting

Calls stop working

Your credits may have run out. Top up in Settings → Billing. If you send many requests at once, you may also be at your plan's requests-a-minute limit; slow down and retry.

My api.kaer.ai key is rejected

Keys from api.kaer.ai do not work on app.kaer.ai/api/v1, and the other way round. Use a key from Settings → Developers here.

Next steps

Connect assistantsUse Kaer tools from Claude, Cursor and other MCP clients.Billing and creditsHow credits and top-ups work.