OpenAI SDKs

Checked against the OpenAI SDKs' docs on 2026-10-11

The official OpenAI SDKs for Python and JavaScript work unchanged: give the client our base URL and your key, and name the model.

Base URL: https://api.creeksidecompute.ai/v1
API: Chat Completions, and Responses without stored state
Model id: qwen/qwen3.8-27b, or qwen/qwen3.6-35b-a3b for speed, or meta/muse-glimmer-30b, which always reasons first
Context window: 131072 tokens, up to 65536 of them output
Key: In CREEKSIDE_API_KEY. Make one at creeksidecompute.ai/account

Set it up

Make a key. Sign in at creeksidecompute.ai/account with your email and make a key. Each request is paid from prepaid credit: top up by card from $10, or redeem a promotional code.
Put the key in an environment variable. Add this line to your shell profile (for example ~/.zshrc), then open a new terminal:
shell
```
export CREEKSIDE_API_KEY=<your key>
```
Install the SDK: pip install openai, or npm install openai.

Point the client at Creekside. In Python:

Python

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.creeksidecompute.ai/v1",
    api_key=os.environ["CREEKSIDE_API_KEY"],
)

response = client.chat.completions.create(
    model="qwen/qwen3.8-27b",
    messages=[{"role": "user", "content": "Explain prefix caching."}],
)
print(response.choices[0].message.content)

In JavaScript or TypeScript:

JavaScript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.creeksidecompute.ai/v1",
  apiKey: process.env.CREEKSIDE_API_KEY,
});

const response = await client.chat.completions.create({
  model: "qwen/qwen3.8-27b",
  messages: [{ role: "user", content: "Explain prefix caching." }],
});
console.log(response.choices[0].message.content);

Streaming works as usual. With stream_options={"include_usage": True} the last chunk carries the token counts, with an empty choices list.

Thinking

Thinking is off unless you ask for it. Pass reasoning_effort: "low", "medium" and "high" turn it on, "none" and "minimal" keep it off. Both SDKs send the field for any model.

The thinking comes back in message.reasoning, or delta.reasoning when streaming, a field the SDK types do not declare. In Python, read it with getattr. In TypeScript, put // @ts-expect-error above the line that reads it. Thinking tokens are billed as output.

Python

response = client.chat.completions.create(
    model="qwen/qwen3.8-27b",
    messages=[{"role": "user", "content": "Is 391 a prime number?"}],
    reasoning_effort="medium",
)
message = response.choices[0].message
print(getattr(message, "reasoning", None))  # the thinking
print(message.content)  # the answer

Muse Glimmer 30B always reasons before it answers, so reasoning_effort sets how long, not whether. Without it, the model reasons briefly, as at "low"; "medium", "high" and "xhigh" make it reason longer, and "none" and "minimal" act as "low". The reasoning counts toward max_tokens, so leave it room.

What does not work

Stored Responses state. client.responses.create works, but nothing is kept between calls: previous_response_id, conversations and background responses are refused. Send the whole conversation in input.
Hosted tools. Web search, file search, code interpreter and image generation run at OpenAI. Your own function tools work.
Other endpoints. Embeddings, images, audio, files, batches and fine-tuning are not served. We serve /v1/models, /v1/chat/completions, /v1/completions and /v1/responses.
Images in messages, on the Qwen models. They are listed for text only. Muse Glimmer 30B reads images sent as image_url parts with base64 data URLs.

Troubleshooting

AuthenticationError: 401 invalid API key: The key is wrong, or CREEKSIDE_API_KEY is not set where the code runs. To test the key alone: curl -s https://api.creeksidecompute.ai/v1/models -H "Authorization: Bearer $CREEKSIDE_API_KEY".
NotFoundError: 404 model … is not served: The model id is wrong. Use qwen/qwen3.8-27b exactly. GET /v1/models lists every id we serve. A 404 on the path itself means an endpoint we do not serve.
BadRequestError: 400 maximum context length is 131072 tokens: The prompt is longer than the model's window. Drop or summarise the oldest turns and send it again.
402 or 429: A 402 means your prepaid credit does not cover the request: add credit at creeksidecompute.ai/account. A 429 means the key's rate, or the number of requests your account may run at once, is used up; agents that work in parallel reach that sooner. Wait and retry. The API reference lists the limits.

Other tools

Sources, read on 2026-10-11: openai-python, openai-node.