OpenAI-Compatible

Drop-in replacement — just change the base URL

Quickstart: OpenAI-Compatible API

Already using the OpenAI SDK? Switch to Archipelag.io by changing two lines — your base URL and API key.

### Set your API key ```bash export ARCHIPELAG_API_KEY="your-api-key-here" ``` ### Use the OpenAI SDK with Archipelag.io {% tabs() %} {% tab(name="Python") %} ```python from openai import OpenAI client = OpenAI( base_url="https://api.archipelag.io/api/v1", api_key="your-api-key-here", # or set ARCHIPELAG_API_KEY ) response = client.chat.completions.create( model="llm-chat", messages=[ {"role": "user", "content": "Say hello in three languages"} ], ) print(response.choices[0].message.content) ```

```javascript import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://api.archipelag.io/api/v1", apiKey: "your-api-key-here", // or set ARCHIPELAG_API_KEY }); const response = await client.chat.completions.create({ model: "llm-chat", messages: [ { role: "user", content: "Say hello in three languages" }, ], }); console.log(response.choices[0].message.content); ```

{% end %}

Run it

python hello.py
# or: node hello.mjs

Hello! (English)
Bonjour! (French)
Hola! (Spanish)

{% end %}

Streaming

Streaming works exactly as you’d expect with the OpenAI SDK:

{% tab(name="Python") %} ```python stream = client.chat.completions.create( model="llm-chat", messages=[ {"role": "user", "content": "Say hello in three languages"} ], stream=True, ) for chunk in stream: content = chunk.choices[0].delta.content if content: print(content, end="", flush=True) ```

```javascript const stream = await client.chat.completions.create({ model: "llm-chat", messages: [ { role: "user", content: "Say hello in three languages" }, ], stream: true, }); for await (const chunk of stream) { const content = chunk.choices[0]?.delta?.content; if (content) process.stdout.write(content); } ```

{% end %}

Model mapping

The API accepts standard OpenAI model names and maps them to Archipelag.io Cargos:

OpenAI model	Archipelag.io Cargo
`gpt-3.5-turbo`	`llm-chat`
`gpt-4`	`llm-chat`
`gpt-4o`	`llm-chat`
`llm-chat`	`llm-chat` (direct)

Note

All model names currently route to the `llm-chat` Cargo (Mistral 7B). As more models become available, each will get its own Cargo slug.

Key differences from OpenAI

Base URL: https://api.archipelag.io/api/v1 (not https://api.openai.com/v1)
API key header: Same format — Authorization: Bearer <key>
Model names: Use llm-chat directly, or standard OpenAI names which are mapped automatically
Rate limits: 100 requests/minute (per API key)

Next steps

curl quickstart — raw HTTP for debugging
Python SDK quickstart — native SDK with streaming iterators
User Guide — models, pricing, error handling