ElevenLabs Text to Speech JavaScript: SDK, Code & Cost

If you searched for “elevenlabs text to speech javascript,” here is the answer: ElevenLabs ships an official TypeScript SDK — the @elevenlabs/elevenlabs npm package — that turns text into speech with a few lines of JavaScript. The SDK install is free; each API call costs credits at one credit per character (checked 4 Nov 2026).

What the ElevenLabs JavaScript SDK does

The official ElevenLabs TypeScript SDK wraps the REST API in typed methods. Instead of hand-building HTTP requests, you call textToSpeech.convert with a voice ID, model, and text, and get audio back. The docs quickstart runs a working example with npx tsx example.mts.

The SDK covers the full API surface — text to speech, voices, speech-to-text, dubbing, and conversational agents. For this post we focus on text to speech, the most common use case.

It works in Node.js 18+ and in modern browsers via bundlers. The package is typed, so editors autocomplete every parameter — voice_id, model_id, output_format, and voice settings all come with types.

The package ships with full TypeScript definitions built in. Every method, option, and response type is declared, so your editor flags mistakes before you run anything. JavaScript users get the same autocomplete through IntelliSense — no separate @types package needed.

Streaming is built in. The SDK can return audio as a stream, so your app can start playing the first chunk before the rest is generated — essential for real-time voice apps and chatbots.

Authentication uses an API key you create in your ElevenLabs account. Keys are scoped per workspace on paid plans, which matters for team projects.

The SDK is maintained by ElevenLabs alongside the API — when new models or voices launch, the package updates rather than leaving you to track endpoint changes.

Compare this with the raw REST API: the SDK saves you from managing audio encoding, retry logic, and parameter validation by hand. For production apps, the SDK is the recommended path.

How to use ElevenLabs text to speech with JavaScript: step by step

Each step notes its cost. Installing and writing code is free — you pay credits only when the API generates audio.

1. Install the package: npm install @elevenlabs/elevenlabs. Works with Node 18+; use a bundler for browser apps. Cost: $0 — installing the SDK is free.
2. Create a free ElevenLabs account and generate an API key in your profile settings. Keep the key server-side — never ship it in client code. Cost: $0 — the Free plan includes 10,000 credits per month.
3. Initialize the client with your key: new ElevenLabsClient({ apiKey: process.env.ELEVENLABS_API_KEY }). Store the key in an environment variable. Cost: $0 — initialization makes no API calls.
4. Convert text to speech: await client.textToSpeech.convert(voiceId, { text: “Hello world”, model_id: “eleven_v4” }). The docs quickstart example.mts shows this exact pattern. Cost: 1 credit per character — a 100-character sentence costs 100 credits.
5. Choose your output format — mp3_44100_128 is the safe default; use pcm formats for raw audio pipelines. Cost: free to change; format does not affect credit cost.
6. Save or stream the result. The SDK returns audio you can write to a file or pipe to a player; use the streaming variant for low-latency playback. Cost: generation already billed in step 4; saving and streaming are free.
7. Handle errors and retries — wrap calls in try/catch, retry on 429s, and log usage so a bug never burns your monthly credits. Cost: $0 for code; only successful generations bill credits.

Pricing and credit costs for API use

API text to speech costs one credit per character — identical to the web app. A 1,000-character article narration costs 1,000 credits. Prices in USD (checked 4 Nov 2026).

A $1 first-month Starter offer was shown until Nov 8 when checked on 4 Nov 2026 (rolling badge) — after that it renews at $6/mo. 30,000 credits generate roughly 30,000 characters of speech.

The Free plan’s 10,000 monthly credits are enough for prototyping — about 200 typical chatbot responses. Our guide to ElevenLabs pricing compares every tier, including API rate limits.

Credits reset monthly and do not roll over. For apps, track usage per endpoint — one runaway loop can burn a month’s credits in minutes, so set usage alerts.

Flash v2.5 costs 50% less per character than v4 per the official models reference — a smart default for high-volume apps where top-tier quality is not required.

Note the Free plan works with the API too — the same 10,000 monthly credits power both web-app and SDK generations, so prototypes cost nothing until you need more volume.

Concurrency limits rise with plan tier. If your app converts many requests in parallel, check the rate limits for your plan before launch — the docs list them per tier.

SDK vs REST API vs web app

Use the JavaScript SDK for production Node and browser apps — typed methods, streaming, and maintained updates. Use the raw REST API only when no SDK exists for your language.

Use the web app for one-off generations, voice testing, and non-technical teammates. Everything the SDK can do starts as a web-app experiment — prototype the voice and settings there first.

Use the Python SDK if your stack is Python — it mirrors the TypeScript SDK’s API. Teams mixing stacks can share voice IDs and settings across both.

One practical difference: the SDK throws typed errors for common failures — invalid keys, exhausted quotas, and rate limits — so your catch blocks can react precisely instead of parsing HTTP status codes by hand.

For conversational agents, the SDK’s real-time support pairs with the Agents platform — but that is a separate product with its own pricing, not plain text to speech.

Our guide to ElevenLabs TTS models explains which model_id to pass — v4 for quality, flash v2.5 for cheap high-volume generation.

Pros and cons

The honest trade-offs of the JavaScript SDK:

Who should NOT use the JavaScript SDK

Skip the SDK if you do not code — the ElevenLabs web app generates the same voices with no setup. Content creators, marketers, and one-off users get everything they need in the browser.

Who it suits instead: web developers adding narration to apps, teams building voice chatbots, indie hackers automating audio content, and agencies generating client audio at scale.

It also suits TypeScript shops that value typed APIs — the SDK’s types catch parameter mistakes before they cost credits.

It also suits anyone prototyping voice features fast — the quickstart example produces real audio within minutes of npm install.

Frequently asked questions

Is there an official ElevenLabs JavaScript SDK?

Yes. ElevenLabs maintains an official TypeScript SDK published as the @elevenlabs/elevenlabs npm package, documented with a quickstart that runs via npx tsx example.mts. It wraps the full REST API in typed methods.

How much does the JavaScript SDK cost?

The SDK itself is free to install. You pay for API usage at one credit per character: the Free plan ($0) includes 10,000 credits monthly; Starter is $6/mo (30,000), Creator $22/mo (121,000), Pro $99/mo (600,000) — checked 4 Nov 2026.

How do I install the ElevenLabs npm package?

Run npm install @elevenlabs/elevenlabs in a Node.js 18+ project. Then create an API key in your ElevenLabs account, initialize ElevenLabsClient with it, and call textToSpeech.convert with a voice ID and text.

Can I stream audio with the JavaScript SDK?

Yes. The SDK supports streaming responses, so playback can start before generation finishes — the right choice for chatbots, voice agents, and any real-time application.

Is the SDK or the REST API better?

The SDK for JavaScript/TypeScript projects — it handles encoding, streaming, and parameter validation for you. Use the raw REST API only in languages without an official SDK.

Can I use the SDK in the browser?

Yes, via a bundler — but never expose your API key in client code. Route browser requests through your own server, which holds the key and calls the SDK.

Sources checked

• ElevenLabs TypeScript SDK quickstart (npm package, example.mts) — https://elevenlabs.io/docs/developers/quickstart — checked 4 Nov 2026.
• ElevenLabs models reference, credit costs — https://elevenlabs.io/docs/overview/models — checked 4 Nov 2026.
• ElevenLabs pricing — https://elevenlabs.io/pricing — checked 4 Nov 2026.