SDK

A TypeScript client for every model.

Typed models, tasks, chat and webhook checks. No dependencies. Works in Node, Bun, Deno and edge runtimes.

npm install tovano

Read the SDK docs View on GitHub

import Tovano from "tovano";

const tovano = new Tovano({ apiKey: process.env.TOVANO_API_KEY });

const task = await tovano.tasks.create({
  model: "image/demo-1",
  input: { prompt: "Blue ceramic mug, morning light", n: 4 },
  customer: "client-0482",
});

const done = await tovano.tasks.wait(task.id);
console.log(done.status, done.charged_micro); // "completed" 80000

What's inside

Everything the API does, typed.

Method	What it does
`models.list()`	Every model with its type, unit and price
`tasks.create()`	Run any image, video, audio or agent model
`tasks.wait(id)`	Poll until a task completes or fails, with a timeout
`chat.completions.create()`	Text models, with streaming
`balance()`	Your balance in micro-dollars
`webhooks.verify()`	Check the `Tovano-Signature` header of a webhook

OpenAI-compatible

Already use an OpenAI client? Keep it.

Text models accept the standard chat completions request, including streaming. Change the base URL and the key.

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

Coming next

On the roadmap

Python SDK (pip install tovano)
MCP server, so coding agents can call any model
Batch tasks for large jobs at a lower unit price
Go client

Want one sooner? Tell us.