Use the OpenAI SDK with CaMeL Hub: One Key for GPT, Claude, Gemini & DeepSeek

The OpenAI SDK (`openai` on PyPI and npm) is the official client library OpenAI publishes for Python, Node.js/TypeScript, and several other languages. It wraps HTTP calls to the Chat Completions and Responses APIs in typed, ergonomic methods — `client.chat.completions.create(...)` instead of hand-built `requests.post()` calls. Because it only needs two settings changed, `base_url` and `api_key`, it's also the fastest way to reach CaMeL Hub: point it at our endpoint and the exact same code that talks to OpenAI's models can call Claude, Gemini, DeepSeek, Grok and 480+ other models through one account, one key, and one bill.

OpenAI SDK ↗

1. Create a CaMeL Hub API key

   Sign in to the CaMeL Hub console at https://api.camel-hub.com/console and open the field named API Keys / Tokens (Console → Token). Click New Token, name it something you'll recognize later (e.g. "openai-sdk-test"), and copy the key right away — like an OpenAI key, it's shown once and should be treated as a secret, never committed to source control.

2. Install the OpenAI SDK

   Install whichever language binding you use. Any current 1.x release works unmodified — CaMeL Hub speaks the exact request/response shape the SDK expects, so there's no compatibility shim or fork to install.

   pip install --upgrade openai
   
   # Node.js / TypeScript instead:
   # npm install openai@latest

3. Point the client at CaMeL Hub

   Create the client with two overrides: `base_url` set to CaMeL Hub's endpoint and `api_key` set to the token from step 1. The SDK appends resource paths such as `/chat/completions` to whatever you pass as `base_url`, so it must already include the `/v1` segment — use `https://api.camel-hub.com/v1` exactly, not the bare `https://api.camel-hub.com` host.

   from openai import OpenAI
   
   client = OpenAI(
       base_url="https://api.camel-hub.com/v1",
       api_key="<YOUR_CAMEL_HUB_KEY>",
   )

4. Send your first request

   Call `chat.completions.create` exactly as you would against OpenAI's own API. Pick any model name from the model marketplace — start with a general-purpose one like `gpt-5.5`.

   completion = client.chat.completions.create(
       model="gpt-5.5",
       messages=[
           {"role": "user", "content": "Explain what CaMeL Hub does in one sentence."}
       ],
   )
   
   print(completion.choices[0].message.content)

5. Call GPT, Claude, Gemini or DeepSeek with the exact same code

   This is the core benefit of routing the OpenAI SDK through CaMeL Hub: the request format never changes, only the `model` string does. CaMeL Hub translates each call into the target provider's native protocol server-side, so a single client instance can round-robin across vendors.

   for model in ["gpt-5.5", "claude-sonnet-5", "gemini-3.1-pro", "deepseek-v4-flash"]:
       resp = client.chat.completions.create(
           model=model,
           messages=[{"role": "user", "content": "Say hello in five words."}],
       )
       print(model, "->", resp.choices[0].message.content)

6. Stream tokens as they arrive

   Set `stream=True` for incremental output — useful for chat UIs or long completions where you don't want to wait for the full response before showing anything.

   stream = client.chat.completions.create(
       model="gpt-5.5",
       messages=[{"role": "user", "content": "Write a haiku about API gateways."}],
       stream=True,
   )
   for chunk in stream:
       delta = chunk.choices[0].delta.content
       if delta:
           print(delta, end="", flush=True)

7. Same pattern in Node.js / TypeScript

   The JavaScript/TypeScript SDK mirrors the Python one field for field. Read the key from an environment variable rather than hardcoding it in the repo.

   import OpenAI from "openai";
   
   const client = new OpenAI({
     baseURL: "https://api.camel-hub.com/v1",
     apiKey: process.env.CAMEL_HUB_API_KEY,
   });
   
   const completion = await client.chat.completions.create({
     model: "gpt-5.5",
     messages: [
       { role: "user", content: "Explain what CaMeL Hub does in one sentence." },
     ],
   });
   
   console.log(completion.choices[0].message.content);

8. Sanity-check with curl if something goes wrong

   If the SDK raises a connection or authentication error, hit the endpoint directly to tell apart a key problem from a client-setup problem. A working curl call with a failing SDK call almost always means an environment variable or base_url typo (see the FAQ below).

   curl https://api.camel-hub.com/v1/chat/completions \
     -H "Authorization: Bearer <YOUR_CAMEL_HUB_KEY>" \
     -H "Content-Type: application/json" \
     -d '{"model": "gpt-5.5", "messages": [{"role": "user", "content": "ping"}]}'

Frequently asked questions