Gradiated API

Connect your existing OpenAI or Anthropic client to affordable inference.

Connect

Every request needs three values:

Base URL: https://api.gradiated.com
API key: a Gradiated key sent from your server
Model: a model from models and pricing

You can use the official OpenAI SDK or Anthropic SDK, or call the HTTP endpoints directly. You won't need to make any code changes to use our inference. The quickstart includes working examples for Python, Node, and cURL.

For Kev decision requests, use the SystemOne SDK with Python, Node, or cURL. The guide covers the TypeSafe SDK used for Jev and the exact Kev model to select on Gradiated. Kev supports only the default service tier.

Service tiers

Gradiated allows you to select the trade-off you want between performance and cost. You can use the service_tier parameter for this.

Use fast for interactive workloads, default for general production traffic, or flex for background agents.

{
  "model": "YOUR_MODEL",
  "service_tier": "default"
}

See service tiers for prices and complete SDK examples.