Skip to main content

Getting Started

Select Inference Endpoint
Choose preferred region.

Get an API Token​

  1. Navigate to the Aurora Portal
  2. Go to Credentials → API Tokens
  3. Click New Token, enter a descriptive name, select execute:inference:requests, and create it
  4. Copy and securely store the token. It is displayed only once

Use this token to authenticate requests to the Inference API. Never commit it to source control. If it is exposed, revoke it in the Aurora Portal and create a new one.

Using the Token​

Include the token in the Authorization header using the Bearer scheme:

Authorization: Bearer <your-token>

Base URL​


Append the endpoint paths shown below to this base URL.

Quick Start​

Store your token in an environment variable:

export AURORA_API_TOKEN="<your-token>"

List the public model catalog (no API token required):

curl /models

With an API token, the same endpoint returns the models available to your tenant:

curl /models \
-H "Authorization: Bearer $AURORA_API_TOKEN"

Then send your first chat completion request:

curl /chat/completions \
-H "Authorization: Bearer $AURORA_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
  "model": "deepseek/deepseek-v4-flash",
  "messages": [
    {"role": "user", "content": "Say hello in one sentence."}
  ]
}'

Stream a response​

Set stream to true to receive text as it is generated. Use curl -N to display chunks without buffering, and request token usage with stream_options.include_usage:

curl -N /chat/completions \
-H "Authorization: Bearer $AURORA_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
  "model": "deepseek/deepseek-v4-flash",
  "messages": [{"role": "user", "content": "Say hello in one sentence."}],
  "stream": true,
  "stream_options": {"include_usage": true}
}'

For the streaming response format and error handling, see Create a chat completion.

Next Steps​