OpenAI SDK
Use the official OpenAI SDK for Python or Node with Nexrouting: set the base URL and your key, and every model works.
La documentazione è in inglese.
The OpenAI SDKs work with Nexrouting as they are. Set the base URL to https://api.nexrouting.com/v1 and use your Nexrouting key. Any model that lists /v1/chat/completions in GET /v1/models works this way, Claude and Gemini models included.
Install
pip install openaiChat completions
import osfrom openai import OpenAI client = OpenAI( base_url="https://api.nexrouting.com/v1", api_key=os.environ["NEXROUTING_API_KEY"],) reply = client.chat.completions.create( model="openai/gpt-6-astra", messages=[{"role": "user", "content": "Hello"}],)print(reply.choices[0].message.content)The base URL also works as https://nexrouting.com/api/v1, for networks that only allow the main domain.
Configure with environment variables
Both SDKs read OPENAI_BASE_URL and OPENAI_API_KEY. With these set, existing code runs on Nexrouting without changes; only the model IDs need the provider/ part.
export OPENAI_BASE_URL="https://api.nexrouting.com/v1"export OPENAI_API_KEY="$NEXROUTING_API_KEY"Streaming
Pass stream=True and read the chunks. Add stream_options to get the token counts in the last chunk; see Streaming.
stream = client.chat.completions.create( model="openai/gpt-6-astra", messages=[{"role": "user", "content": "Write a haiku about routing."}], stream=True, stream_options={"include_usage": True},)for chunk in stream: if chunk.choices and chunk.choices[0].delta.content: print(chunk.choices[0].delta.content, end="") if chunk.usage: print("\n", chunk.usage)Responses API
Models that list /v1/responses in their endpoints accept the Responses API. A few models, such as some OpenAI pro models, are only available there.
response = client.responses.create( model="openai/gpt-6-astra", input="Explain routing in one sentence.",)print(response.output_text)Embeddings
result = client.embeddings.create( model="openai/text-embedding-3-small", input=["first text", "second text"],)print(len(result.data[0].embedding))List the models
client.models.list() calls GET /v1/models and works as usual. Each entry has extra fields such as pricing, lane and endpoints; see List models.
Errors and retries
Errors use the OpenAI shape, so the SDK raises its usual exceptions: AuthenticationError for 401, PermissionDeniedError for 403, NotFoundError for 404, RateLimitError for 429 and APIStatusError for the rest, such as 402 insufficient_credits. The SDK retries 429 and 5xx responses by itself. See Errors.
Not available
Files, batches, fine-tuning, assistants, moderations, image generation, audio and realtime are not available. Those paths return 404 not_found.