Skip to main content
OrcaRouter exposes a first-class OpenAI surface. Point your existing OpenAI SDK at OrcaRouter’s base URL and your code keeps working the way your SDK expects — streaming, tool calling, structured outputs (json_object and json_schema), and vision are all supported. When you target a model from a different provider, OrcaRouter’s translation layer adapts request and response to that provider’s native shape so your SDK code stays unchanged.

​
Python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="sk-orca-...",
)

​
TypeScript / Node

import OpenAI from "openai";

const openai = new OpenAI({
  baseURL: "https://api.orcarouter.ai/v1",
  apiKey: "sk-orca-...",
});

​
Async Python

from openai import AsyncOpenAI

client = AsyncOpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="sk-orca-...",
)

​
Using environment variables

The OpenAI SDK reads OPENAI_API_KEY and OPENAI_BASE_URL by default. Set them once and the SDK picks them up without per-call config: 
export OPENAI_API_KEY="sk-orca-..."
export OPENAI_BASE_URL="https://api.orcarouter.ai/v1"

​
What changes in your code

Only the base URL and the API key. Request parameters, response shape, streaming protocol, error handling — all unchanged. Model names are provider-prefixed (openai/gpt-4o-mini, anthropic/claude-sonnet-4.6, google/gemini-2.5-pro, deepseek/deepseek-chat, grok/grok-4-fast-reasoning, qwen/qwen3.6-plus, kimi/kimi-k2.6, minimax/minimax-m2.7) so clients always know which provider served the request; OrcaRouter handles the cross-provider translation internally through the same client object. Web search works with the OpenAI SDK against OrcaRouter. See Advanced / Web search for the per-endpoint parameter shapes and which models support it.

​
Other SDKs

Using the Anthropic SDK or google-genai SDK directly? See Compatibility / Anthropic SDK and Compatibility / Google GenAI SDK.