Quickstart

Create an API Key

All requests to the Ollm API require authentication using an API key.

Create a key from the dashboard

1. Open the Dashboard in the Ollm console.
2. Navigate to API Keys.
3. Click Generate Key.
4. Enter a descriptive name (for example, production-key or local-dev).
5. Click Create Key.

Your API key will be shown once. Copy it and store it securely.

Important Keep your API key secret. Do not expose it in frontend code or commit it to version control.

Authenticate Requests

Ollm uses Bearer token authentication. Send your API key in the Authorization header.

Required headers

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Make Your First Request

Ollm is OpenAI-compatible. You explicitly choose the model in your request.

Example: Chat completion (curl)

curl -X POST https://api.ollm.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "phala/gemma-3-27b-it",
    "messages": [
      { "role": "user", "content": "Hello" }
    ]
  }'

Example: Chat completion (JavaScript)

const res = await fetch("https://api.ollm.com/v1/chat/completions", {
  method: "POST",
  headers: {
    Authorization: "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "phala/gemma-3-27b-it",
    messages: [{ role: "user", content: "Hello" }],
  }),
});
const data = await res.json();
console.log(data);

If the request is successful, you’ll receive a model response along with metadata tied to that specific request.

View and Verify the Request

After sending a request, you can inspect it directly in the Ollm dashboard.

From the Messages view, you can:

• See request status (Success, Failed, Pending)
• View the selected model and provider
• Confirm TEE attestation status (Verified / Pending / Failed)
• Inspect cryptographic verification details

What you can see

• Request ID
• Model and provider
• Token usage, latency, and cost
• Attestation status
• Cryptographic hashes and signatures

What you cannot see

• Prompt contents
• Model outputs
• Any plaintext inference data

This design ensures transparency and auditability without exposing sensitive data.

Understand Attestation Status

Each request includes a TEE attestation result:

• Verified The request was executed inside a verified Trusted Execution Environment. Cryptographic proof is available.
• Pending Attestation is still being finalized.
• Failed The request did not meet verification requirements.

Attestation details can be independently verified using the provided hashes and signatures.