Is ElevenLabs free?

How to prepare your first personal narration

Start with one short paragraph you wrote yourself. A personal listening exercise lets you learn the controls without committing the whole monthly allowance to a long script. These steps describe the documented workflow; we have not completed an account test.

1. Check the account before creating

   Sign in, inspect Subscription and confirm that you are using Free rather than a trial or paid plan. Note the remaining balance and next reset. Open Text to Speech for this exercise; other products have different costs and export rules.

2. Choose an available voice and model

   Select a default voice that suits the script and a model actually selectable on your account. Multilingual v2 is a possible starting point if available. A model appearing in public documentation does not prove it is included for your account. Stop at an upgrade gate rather than assuming payment is required for every route.

3. Finish the text before generating

   Paste the paragraph below, then check punctuation, spelling and paragraph breaks. Keep a plain-text copy. Leave unfamiliar audio tags or pronunciation dictionaries out of the first attempt so you can identify what changed later.

4. Read the displayed cost

   Record the voice, model, settings, character count and credit cost before selecting Generate. Changing the model or voice can change the charge. Keep enough allowance for a correction instead of assuming an unsatisfactory first take will always have a free retry.

5. Generate once and listen carefully

   Listen from beginning to end. Check that every sentence is present, names are understandable, pauses sound natural and the final word is not cut off. Save a usable take before editing the text or making another request.

6. Save the file and your notes

   Use the available download action, then play the saved file in your own audio player. Record the actual charge and duration alongside the text and settings. Keep a personal practice file private unless the intended sharing meets the applicable usage terms.

Practice: narrate a tiny observatory waking before dawn

This original paragraph gives you a concrete listening task: a character name, a sequence of actions and a gentle change of mood. It is intended for personal practice. No generated recording accompanies it, and a specific voice's pronunciation or delivery is not promised.

The exact text below contains 341 characters, including spaces and punctuation, and 64 words. Paste it as one paragraph without adding quotation marks or tags, then check the count shown by the service.

• Can you follow the scene without looking at the script? Listen for a useful pause between the cloudy sky and the star appearing.
• Check how the voice says Mira and observatory. If a word needs correction, change one thing and review the next request's cost.
• Listen on headphones and a phone speaker. Note clipping, unusual breaths or rushed words rather than describing a voice as universally realistic.
• Keep the exact text, voice, model, settings and saved take together. Identical settings need not produce identical audio on a later generation.
• Do not attach a claimed generation time, audio duration, download quality or watermark result to this exercise until you have checked a real output.

How far does the free narration allowance go?

Free includes 10,000 Creative credits each month, shared across products. Using another Creative tool leaves fewer credits for narration. Unused Free credits do not accumulate, so check your account's reset date rather than assuming a refill on the first day of the calendar month.

A credit is not a word or a fixed second of speech. At a displayed rate of one credit per character, this 341-character paragraph would cost 341 credits for one generation. That is a conditional calculation, not an observed charge. Read the actual model and voice cost before using it as your budget.

API Flash/Turbo discounts do not establish half-price web generation. Keep retries in the budget, and avoid converting the whole balance into an exact number of finished minutes: delivery speed and discarded takes change the result.

Keep the first request comfortably below the limit

The checked Help Center states a 2,500-character limit per Free web Text to Speech request. Our practice paragraph fits comfortably within it. That request ceiling is separate from the monthly allowance.

Longer model-specific API limits are not the Free web limit. If your script is too long for the displayed controls, split it at natural sentence or paragraph boundaries and keep voice/settings notes for consistency. Each new generation can use allowance; a longer script does not create extra credits.

When is another take actually free?

For the generic web TTS workflow, ElevenLabs documents two free regenerations when text, voice and model remain unchanged, the first take is less than two hours old, and you have not left or refreshed the page. Voice-setting sliders may change under that rule. Check the Regenerate indicator before clicking.

Eleven v3 behaves differently: a Generate click returns two alternatives while charging for one generation. Clicking Generate again is another charge. Two alternatives from one click do not mean a second click is free.

The API does not include these web free regenerations. If you edit the script, switch the model or cannot see a free retry clearly indicated, budget for a new charge instead of repeatedly clicking until the voice sounds right.

Download the narration and inspect the saved file

The documented TTS download action appears after generation. Previous takes can also be downloaded from History. On a narrow screen, the History icon is above Generate speech; on a wider screen, look for the History tab in the right panel.

Help describes MP3 and WAV options, with additional formats under Advanced. Check what your actual Free account offers. Choosing WAV does not, by itself, prove a higher-quality source recording, and documentation describing a format is not our verification of a successful Free download.

Play the saved file outside the browser before planning further edits. Check the beginning, ending, audible branding and any metadata you can inspect. We have not downloaded an ElevenLabs output, so audio watermark and export quality remain untested here.

Can Free narration go into a monetized video or client project?

Free output is for non-commercial use. For eligible non-commercial sharing of TTS content, ElevenLabs requires elevenlabs.io or 11.ai in the title. Adding that credit does not grant commercial permission for a monetized video, advertisement or client deliverable.

A later paid upgrade does not retrospectively license recordings made on Free. The publishing help distinguishes content generated during a paid subscription, subject to input rights, service-specific terms and Beta restrictions. Choose the appropriate licensed workflow before producing a paid project's final narration.

We provide a script and documented steps rather than a Free-generated audio showcase. Permission to host a Free recording on an ad-supported directory has not been established. Keep your personal test private until its intended distribution is covered.

Treat developer access as a separate budgeting exercise

Use the web playground if your immediate goal is hearing and saving one paragraph. API access is useful when an application must generate speech programmatically, but you do not need an API key for this personal web exercise.

The current API pricing page describes dollar metering rather than Creative credits, and groups Free with Pay As You Go in its comparison. Its model allowances should not be added together as separate free pools or treated as proof of the same debit rate as the website. Inspect the account's developer usage and applicable model pricing before making requests.

Keep any API key server-side or in a private local environment. Downloaded API audio still has applicable usage restrictions; choosing an API route does not remove the Free commercial-use rule or create free retries.

When to choose Cartesia instead

Compare Cartesia when you want another browser-based voice workflow. Its official pricing lists Sonic-3.6 and a Free plan with 20,000 monthly credits. Use the Playground and confirm the model available to your Free account before trying the same original paragraph.

The Free allowance is shared between speech products, and a commercial-use license starts on a paid plan. We have not tested Cartesia's Free model controls or compared its narration quality with ElevenLabs. A newer model name alone does not establish better delivery or suitable output rights.