@cartesia/cartesia-js

Cartesia TypeScript Library

The Cartesia TypeScript library provides convenient access to the Cartesia API from TypeScript.

Installation

npm i -s @cartesia/cartesia-js

Reference

A full reference for this library is available here.

Usage

Instantiation

Instantiate and use the client with the following:

import { CartesiaClient } from "@cartesia/cartesia-js";

const client = new CartesiaClient({ apiKey: "YOUR_API_KEY_HEADER" });
await client.tts.bytes({
    modelId: "sonic-english",
    transcript: "Hello, world!",
    voice: {
        mode: "id",
        id: "694f9389-aac1-45b6-b726-9d9369183238",
    },
    language: "en",
    outputFormat: {
        container: "raw",
        sampleRate: 44100,
        encoding: "pcm_f32le",
    },
});

TTS over WebSocket

import { CartesiaClient } from "@cartesia/cartesia-js";

const cartesia = new CartesiaClient({
    apiKey: "your-api-key",
});

// Initialize the WebSocket. Make sure the output format you specify is supported.
const websocket = cartesia.tts.websocket({
    container: "raw",
    encoding: "pcm_f32le",
    sampleRate: 44100,
});

try {
    await websocket.connect({
        // If using Node.js, you can pass a custom WebSocket constructor, such as from `ws`.
        // This is not needed for browser usage, so you can call connect() without any arguments.
        WebSocket: WS,
    });
} catch (error) {
    console.error(`Failed to connect to Cartesia: ${error}`);
}

// Create a stream.
const response = await websocket.send({
    modelId: "sonic-english",
    voice: {
        mode: "id",
        id: "a0e99841-438c-4a64-b679-ae501e7d6091",
    },
    transcript: "Hello, world!",
    // The WebSocket sets output_format on your behalf.
});

// Access the raw messages from the WebSocket.
response.on("message", (message) => {
    // Raw message.
    console.log("Received message:", message);
});

// You can also access messages using a for-await-of loop.
for await (const message of response.events("message")) {
    // Raw message.
    console.log("Received message:", message);
}

Input Streaming with Contexts

const contextOptions = {
    contextId: "my-context",
    modelId: "sonic-english",
    voice: {
        mode: "id",
        id: "a0e99841-438c-4a64-b679-ae501e7d6091",
    },
};

// Initial request on the context uses websocket.send().
// This response object will aggregate the results of all the inputs sent on the context.
const response = await websocket.send({
    ...contextOptions,
    transcript: "Hello, world!",
});

// Subsequent requests on the same context use websocket.continue().
await websocket.continue({
    ...contextOptions,
    transcript: " How are you today?",
});

See the input streaming docs for more information.

Playing audio in the browser

(The WebPlayer class only supports playing audio in the browser and the raw PCM format with fp32le encoding.)

// If you're using the client in the browser, you can control audio playback using our WebPlayer:
import { WebPlayer } from "@cartesia/cartesia-js";

console.log("Playing stream...");

// Create a Player object.
const player = new WebPlayer();

// Play the audio. (`response` includes a custom Source object that the Player can play.)
// The call resolves when the audio finishes playing.
await player.play(response.source);

console.log("Done playing.");

Request And Response Types

The SDK exports all request and response types as TypeScript interfaces. Simply import them with the following namespace:

import { Cartesia } from "@cartesia/cartesia-js";

const request: Cartesia.VoiceChangerBytesRequest = {
    ...
};

Exception Handling

When the API returns a non-success status code (4xx or 5xx response), a subclass of the following error will be thrown.

import { CartesiaError } from "@cartesia/cartesia-js";

try {
    await client.tts.bytes(...);
} catch (err) {
    if (err instanceof CartesiaError) {
        console.log(err.statusCode);
        console.log(err.message);
        console.log(err.body);
    }
}

Advanced

Retries

The SDK is instrumented with automatic retries with exponential backoff. A request will be retried as long as the request is deemed retriable and the number of retry attempts has not grown larger than the configured retry limit (default: 2).

A request is deemed retriable when any of the following HTTP status codes is returned:

• 408 (Timeout)
• 429 (Too Many Requests)
• 5XX (Internal Server Errors)

Use the maxRetries request option to configure this behavior.

const response = await client.tts.bytes(..., {
    maxRetries: 0 // override maxRetries at the request level
});

Timeouts

The SDK defaults to a 60 second timeout. Use the timeoutInSeconds option to configure this behavior.

const response = await client.tts.bytes(..., {
    timeoutInSeconds: 30 // override timeout to 30s
});

Aborting Requests

The SDK allows users to abort requests at any point by passing in an abort signal.

const controller = new AbortController();
const response = await client.tts.bytes(..., {
    abortSignal: controller.signal
});
controller.abort(); // aborts the request

Runtime Compatibility

The SDK defaults to node-fetch but will use the global fetch client if present. The SDK works in the following runtimes:

• Node.js 18+
• Vercel
• Cloudflare Workers
• Deno v1.25+
• Bun 1.0+
• React Native

Customizing Fetch Client

The SDK provides a way for your to customize the underlying HTTP client / Fetch function. If you're running in an unsupported environment, this provides a way for you to break glass and ensure the SDK works.

import { CartesiaClient } from "@cartesia/cartesia-js";

const client = new CartesiaClient({
    ...
    fetcher: // provide your implementation here
});

Contributing

While we value open-source contributions to this SDK, this library is generated programmatically. Additions made directly to this library would have to be moved over to our generation code, otherwise they would be overwritten upon the next generated release. Feel free to open a PR as a proof of concept, but know that we will not be able to merge it as-is. We suggest opening an issue first to discuss with us!

On the other hand, contributions to the README are always very welcome!