@zauto/canopus NPM

Canopus SDK

The Canopus SDK provides a client for interacting with the Canopus AI platform, allowing developers to work with various AI models such as Text, Speech-to-Text (STT), Text-to-Speech (TTS), and Embedding models via a simple JavaScript API.

Installation

You can install the Canopus SDK using npm:

npm install @zauto/canopus

or using yarn:

yarn add @zauto/canopus

Getting Started

Import the SDK

import {Canopus} from '@zauto/canopus';

Create an Instance of CanopusSDK

To use the SDK, instantiate it with your API key:

const canopus = new Canopus({ apiKey: 'YOUR_API_KEY' });

API Methods

1. Verify API Key

Verify if the provided API key is valid.

try {
  await canopus.verifyApiKey();
  console.log('API key verified successfully.');
} catch (error) {
  console.error('Invalid API key:', error.message);
}

2. Get Models

Retrieve the list of available models. Optionally, filter models by type (TEXT, STT, TTS, EMBEDDING).

try {
  const models = await canopus.getModels(); // Get all models
  const textModels = await canopus.getModels('TEXT'); // Get models of type 'TEXT'
  console.log(models);
} catch (error) {
  console.error('Error fetching models:', error.message);
}

3. Call Text Models

Call a text model with a prompt and optional chat history or system prompt.

try {
  const response = await canopus.callTextModels('MODEL_ID', 'What is the capital of France?');
  console.log(response.data);
} catch (error) {
  console.error('Error calling text model:', error.message);
}

4. Call TTS Model

Convert text to speech using a TTS model.

try {
  const response = await canopus.callTtsModel('MODEL_ID', 'Hello, how are you?', 'en-IN', 'meera');
  console.log(response.data);
} catch (error) {
  console.error('Error calling TTS model:', error.message);
}

5. Call STT Model

Convert an audio file to text using an STT model.

try {
  const response = await canopus.callSttModel('MODEL_ID', '/path/to/audio.mp3', 'Optional prompt');
  console.log(response.data);
} catch (error) {
  console.error('Error calling STT model:', error.message);
}

6. Call Embedding Model

Generate embeddings for an array of strings using an Embedding model.

try {
  const response = await canopus.callEmbeddingModel('MODEL_ID', ['sentence 1', 'sentence 2']);
  console.log(response.data);
} catch (error) {
  console.error('Error calling embedding model:', error.message);
}