Getting started

You'll need a TensorLoop account and a terminal. The whole flow takes about five minutes.

1. Sign up

Head to tensorloop.tech/auth/sign-up and create an account with email + password (or Google, if your deployment has it wired up).

You're automatically provisioned on the free plan, which gives you:

• Up to 2 API keys
• A $5 rolling 30-day budget per key
• 30 requests per minute per key
• Access to gpt-4o-mini

See Models for what changes on pro.

2. Mint an API key

1. Open tensorloop.tech/dashboard.
2. In the Create key panel, fill in:
   • Alias — a short name, e.g. my-laptop. Lowercase letters, numbers, dashes.
   • Models — select one or more from your plan.
   • Budget, rate limit, expiry — defaults are sane.
3. Click Create API key.

The key appears in a dialog once. Copy it to a password manager — it is not retrievable afterward.

3. Make a request

TensorLoop is OpenAI-compatible. Point your client at https://litellm.tensorloop.tech/v1:

curl https://litellm.tensorloop.tech/v1/chat/completions \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

You'll get a standard OpenAI-shaped completion:

{
  "id": "chatcmpl-...",
  "model": "gpt-4o-mini",
  "choices": [
    {
      "index": 0,
      "message": {"role": "assistant", "content": "Hello, friend."},
      "finish_reason": "stop"
    }
  ],
  "usage": {"prompt_tokens": 12, "completion_tokens": 5, "total_tokens": 17}
}

4. Watch it land in the dashboard

Refresh the dashboard. Within a few seconds you'll see:

• The call in Activity → Recent calls with tokens and cost.
• A new bar in Analytics → Daily spend (30d).
• An updated Spend column for the key.

What's next

• Read Authentication for the details on key scopes, expiry, and rotation.
• Read the Chat completions reference for every supported parameter.
• See Examples for snippets in curl, Python, and JavaScript.