Gateway Integrations

ORGN Gateway exposes an OpenAI-compatible API at https://api.gateway.orgn.com/v1. Existing applications, SDKs, and tools that speak the OpenAI wire format can route inference through Gateway with minimal changes — swap the base URL, add your sk-ollm-* API key, and select a model.

Authentication is always Authorization: Bearer sk-ollm-YOUR_API_KEY. Model IDs use underscores (near_qwen3_30b, vercel_claude_sonnet_4_6) — not slashes.

Integration options

All approaches hit the same endpoint and support the same model catalog. The model you select determines execution security — TEE or ZDR — not the client library.

Core endpoints

See Response anatomy for the full request and response contract, including attestation metadata on TEE model responses.

Minimal integration

curl -X POST https://api.gateway.orgn.com/v1/chat/completions \
  -H "Authorization: Bearer sk-ollm-YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "phala_deepseek_r1",
    "messages": [
      { "role": "user", "content": "Hello" }
    ]
  }'

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.OLLM_API_KEY,       // sk-ollm-...
  baseURL: "https://api.gateway.orgn.com/v1",
});

const completion = await client.chat.completions.create({
  model: "near_qwen3_30b",
  messages: [{ role: "user", content: "Hello" }],
});

Model selection and execution types

Gateway does not auto-route models. The model field in your request explicitly selects the execution environment:

Discover models at runtime rather than hardcoding IDs:

curl https://api.gateway.orgn.com/v1/models \
  -H "Authorization: Bearer sk-ollm-YOUR_API_KEY"

Or browse the catalog in the Gateway console Model List.

IDE and tool configuration

Any tool that supports a custom OpenAI-compatible endpoint can use Gateway:

Create a separate API key per environment (development, staging, production) to simplify rotation.

Verifying TEE inference

After integrating a TEE model, verify attestation independently:

1. Send a request through your integration.
2. Open scanner.orgn.com and locate the request.
3. Inspect attestation status (Verified, Pending, or Failed) and cryptographic signatures.

Verification happens in Scanner, not the Gateway console or your application client. See Verification for the attestation flow.

Guides

Step-by-step migration and integration guides for common stacks:

What's next

• Vercel AI SDK — @orgn/gateway provider setup
• Response anatomy — request and response structure
• Authentication — API keys and SSO
• API reference — full OpenAPI specification