Use cubic in your own products

Base URL

https://api.squarecloud.app/v2/ai

API key

Your account's key, with the ai:chat scope

Model

cubic

import OpenAI from "openai";const client = new OpenAI({  baseURL: "https://api.squarecloud.app/v2/ai",  apiKey: process.env.SQUARECLOUD_API_KEY,});const completion = await client.chat.completions.create({  model: "cubic",  messages: [    { role: "system", content: "You are a helpful assistant." },    { role: "user", content: "Explain what Square Cloud is in one sentence." },  ],});console.log(completion.choices[0].message.content);

curl https://api.squarecloud.app/v2/ai/chat/completions \  -H "Authorization: Bearer $SQUARECLOUD_API_KEY" \  -H "Content-Type: application/json" \  -d '{    "model": "cubic",    "messages": [      { "role": "user", "content": "Explain what Square Cloud is in one sentence." }    ]  }'

• OpenAI-compatible

  Any OpenAI SDK works once its base URL points at api.squarecloud.app/v2/ai. Errors come back in the OpenAI format too.

• Function calling

  Declare up to 32 tools in the OpenAI format. The model decides when to call one, and your code runs it and sends back the result.

• Built-in web search

  When a question needs current information, the model searches the web on its own and returns only the final answer, citing the source URLs.

• Limits of its own

  The gateway's token limits are separate from the dashboard assistant's: API traffic never uses up what Cubic has left in the dashboard, and vice versa.

On Enterprise, the token limit grows with the plan size. Context and output apply per request.

• Text only

  Text in, text out, by design. Images, audio or files return a 400. Cubic, in the dashboard, is a separate product and accepts images and voice.

• No memory between calls

  Like OpenAI, the API does not keep conversations. Send the history in messages on every request, up to 100 messages.

• Full response, no streaming

  For now the response arrives in one piece. Requests with "stream": true return a 400.