Build the idea.
Delegate the rest.
Connect intelligence, projects, and files with a focused set of APIs. From your first request to a complete agent workflow.
Claude + Gemini · Streaming · Tool calls
POST /v1/chat/completions
// One familiar interface. Your choice of model.
{
"model": "claude-sonnet-4-6",
"messages": [{
"role": "user",
"content": "Let's build something useful."
}],
"stream": true
}The building blocks
Three services. One workflow.
Choose the service you need. Each reference includes request shapes, responses, and authentication.
AI Gateway
Talk to Claude and Gemini, stream responses, call tools, and generate embeddings or images.
Explore the gatewayPlatform API
Bring projects, tasks, conversations, and account data into your application workflows.
Explore the platformFile Service
Upload, organize, and manage user files, with a trash and restore flow built into the API.
Explore file storageFrom zero to first response
A small integration.
A much bigger toolkit.
Use the OpenAI clients you already know for the gateway, or start with the Go SDK guide. Keep your model choice flexible.
Read the quickstart- Choose your connectionUse a provisioned gateway key on your server. Browser apps connect through the authenticated platform proxy.
- Pick a client and modelFollow a language guide and choose Claude or Gemini. The model name selects the provider.
- Make it your workflowAdd streaming, tool calls, and your project data. Read the operations guide for request IDs and errors.
Start in your language
Your stack. Your next step.
Practical examples for model discovery, streaming responses, and embeddings.