Frequently Asked Questions About Elevenlabs

The most important answers about ElevenLabs

ElevenLabs is an AI platform that lets you convert text to speech, clone voices, translate recordings, and edit audio. You can try it for free, but for commercial use, you’ll typically need a paid subscription; be sure to check the terms and conditions for the specific feature you’re using.

What is ElevenLabs and what is it used for?

What is ElevenLabs?

ElevenLabs is an AI platform for creating and processing speech and other audio. You can have text read aloud, choose a suitable narrator’s voice, or work with an authorized voice clone. Additionally, there are features for transcription, dubbing, sound effects, and putting together larger audio projects.

What can you create with ElevenLabs?

Think of voice-overs for explanatory videos, narrated courses, podcasts, audio stories, and multilingual product presentations. The features you can use depend on your subscription. For a simple explanatory video, a good script and a suitable voice are often all you need. You don’t necessarily need your own voice clone for that.

What does text-to-speech mean?

With text-to-speech, ElevenLabs converts written text into spoken audio. You choose a voice and a suitable model and adjust the settings as needed. You can listen to the result and process it further. The quality also depends on how naturally your text is written.

Is ElevenLabs the same as a voice changer?

No. Text-to-speech generates audio from written text. A voice changer, on the other hand, uses an existing recording and alters the voice in it. ElevenLabs offers both options. If you want to record the pace and emotion yourself, Voice Changer might be a good fit. If you don’t want to record anything at all, text-to-speech is a better choice.

Do you need to record your own voice?

No. You can work with available voices or product type a voice using the features provided for that purpose. Recording is especially relevant if you want to clone your own voice or use an existing recording. Test an available voice first before spending time on an extensive recording session.

Is ElevenLabs suitable for beginners?

Yes. Start with a short text excerpt, choose a voice, and listen to the result. Then change one element at a time—for example, the speed or phrasing—to see what makes the biggest difference. Starting with a full audiobook as your first project makes it harder to spot mistakes and avoid unnecessary costs.

Get started with ElevenLabs

How do you get started with ElevenLabs?

Create an account on the official website and open the desired feature, such as Text to Speech. Choose a voice and a model that supports your language. Enter a short, verified text and generate a sample. Evaluate pronunciation, pace, and intonation before processing a longer text.

Do you need to install a program?

You don’t need to install a separate desktop program to use the web interface. You can work directly through your browser. If you’re using apps or links, check who the provider is and what features are available. Access via another application may involve different costs or settings than using ElevenLabs directly.

Can you use ElevenLabs on Windows , and Mac?

Yes, you can use the website on Windows using a compatible browser on Windows and Mac. Keep your browser up to date and ensure you have a stable connection. To record your voice, you’ll also need a working microphone. A microphone isn’t required for standard text-to-speech.

Do you need a powerful computer?

The service handles the online generation process. You don’t need a high-end local graphics card for this. A well-functioning browser and a reliable internet connection are more important. If you then combine the audio with video or perform complex editing, the platform requirements of your editing software will also apply.

Can you use ElevenLabs offline?

You’ll need an internet connection to use the online generative features. However, you can listen to and edit downloaded audio files offline afterward. Be sure to save your final files outside of your account as well. This makes it easier to transfer them to a client and prevents you from having to rely on the website every time you want to play them.

Dutch voices and natural pronunciation

Does ElevenLabs support Dutch?

Yes, Dutch is supported by various ElevenLabs models. Be sure to check the language support for the feature you’re using. A text-to-speech model and a dubbing tool may have different capabilities. Always test with your own text, especially if it contains brand names, technical terms, or place names.

Could you use a Flemish voice?

Find a voice whose pronunciation and accent match your target audience, and listen to a sample recording of your text. Support for Dutch doesn’t mean every voice automatically sounds Flemish. Pay attention to vowels, intonation, and word choice. Appropriate phrasing is just as important as the choice of speaker.

How do you choose a good voice for an instructional video?

Pay close attention to clarity, calmness, and credibility. An expressive voice isn’t suitable for every instruction. Test a short clip using real product names and a specific action. This way, you can hear whether the voice is suitable for the content itself, rather than just for an appealing demo clip.

Why does an AI voice sometimes sound unnatural?

The issue may lie with the text, voice, model, or settings. Long sentences and a string of abbreviations often sound less pleasant. Read the script aloud yourself and simplify difficult passages. Only then should you adjust the settings. Clear spoken text is a stronger starting point than a formal text full of subordinate clauses.

How can you improve the pronunciation of a brand name?

First, do a quick test using the name in a complete sentence. Where supported, use a pronunciation dictionary or pronunciation guide. The available options vary by model. An alternative spelling in the voice recording script can also help, but keep it separate from text that will be published publicly.

How do you correctly pronounce amounts, dates, and abbreviations?

Write them out in your audio script exactly as they should sound. For example, “twenty-five euros” leaves less room for interpretation than a concise price notation using numbers. Check product versions and codes individually. The correct pronunciation depends on the context; a number can represent an amount, a year, or a model number.

Can you adjust the speech rate?

Yes, supported text-to-speech features have a speed setting. Use it sparingly: too fast a pace reduces intelligibility. If your voice-over doesn’t fit within the video, shorten the script first. A concise explanation usually sounds better than a lot of information spoken at an accelerated pace.

Can you influence emotion and intonation?

This can be done via text, voice commands, and the capabilities of the chosen model. Not every model supports the same instructions. Start by creating a natural script and then add specific instructions. Check whether the emotion matches the message; a cheerful tone may sound inappropriate for serious information.

How to Create a Good Voice-Over

How do you write text that sounds natural when read aloud?

Write for the ear: keep your thoughts brief, use clear verbs, and choose familiar words. Don’t use every line of on-screen text verbatim as voice-over. Focus on what the viewer needs to understand or do. Read the text aloud to yourself once; this will help you spot awkward sentences more easily.

Example of a short instructional text

Open your computer’s settings. Then select Platform and click Activation. Here you can see if “ Windows ” is active. Would you like to enter a different product key? If so, select “Change Product Key” and follow the on-screen steps.

Should you include voice instructions in the text box?

That depends on the model and the feature being used. Some prompts are supported, while others can be spoken verbatim. Therefore, check the model manual. Keep your spoken text and production notes separate, unless the feature explicitly provides a way to interpret those notes.

Can you record long texts in one go?

That depends on the feature and the input limit. For longer productions, a project environment like Studio is often more organized. Divide the content into logical sections. This makes it easier to check or replace a passage without having to recreate the entire story.

How do you maintain consistency throughout a long recording?

Use the same voice, the same model, and similar settings. Stick to a consistent spelling for names and technical terms. Also listen to the transitions between segments. Two separate passages may each sound good on their own, but when placed side by side, they may differ in energy, tempo, or volume.

Can you just rewrite the incorrect sentence?

In appropriate project functions, you can regenerate selected text. Then check how it fits with the surrounding audio. A brand new version may sound different in terms of rhythm or emphasis. Keep the genuine recording until you’re sure the replacement fits better into the whole.

How can you avoid unnecessary credit usage?

Finalize your script before generating a lot of audio. Test difficult names and the selected voice with a short clip. Then generate audio for each logical section. Each complete brand new attempt may incur a usage charge, although some features offer limited options for free regeneration under certain conditions.

Voice Cloning with ElevenLabs

What is voice cloning?

In voice cloning, ElevenLabs uses voice recordings to digitally replicate a voice. This allows you to have brand new texts spoken. The result depends on the source material, method, and model. A voice clone does not guarantee that every emotion, accent, and pronunciation will exactly match those of the genuine speaker.

What is the difference between Instant and Professional Voice Cloning?

Instant Voice Cloning works quickly with short recordings; Professional Voice Cloning uses more extensive material and a separate training process. The professional method requires more preparation and verification. Choose based on your goal and first test whether the simpler method delivers sufficient quality.

How much audio do you need to clone your voice?

For Instant Voice Cloning, ElevenLabs requires approximately one to two minutes of good-quality audio. Professional Voice Cloning requires significantly more material—typically dozens of minutes. Quality is a major factor: a long recording with echo and varying microphone distances is not a good substitute for consistent, clear source material.

Can you create a Professional Voice Clone of someone else?

No. ElevenLabs only allows Professional Voice Cloning for your own voice, with verification. If someone wants to share their professional voice clone with you, that person must create and verify the clone themselves. After that, sharing can take place using the available options.

Do you need permission to clone a voice?

You must have the necessary rights and permission and comply with the rules governing the voice cloning feature you’re using. Finding a recording online does not grant you permission to imitate that voice. Furthermore, do not use voice cloning to mislead people about who actually said something.

How do you make good recordings for a voice clone?

Record in a quiet room with minimal reverberation and maintain a consistent distance from the microphone. Use a single speaker, without background music. Speak in a manner appropriate for the intended end use. Significant variations in recording quality or excessive post-processing can make the result less predictable.

Is your voice clone ready right away?

An Instant Voice Clone can be available quickly. A Professional Voice Clone requires processing and training, which takes longer. Check the status in your account and avoid scheduling important production work for a guaranteed delivery time. Queues, verification, and processing may affect availability.

Dubbing, transcription, and other audio features

Can ElevenLabs dub a video into another language?

Yes. With Dubbing, you can have audio or video translated and re-recorded in supported languages. The platform attempts to preserve the characteristics of the genuine speakers. Check the translation, pronunciation, and timing. A technically successful dubbing may still contain an incorrect meaning or unnatural phrasing.

Is dubbing the same as subtitling?

No. Dubbing replaces or adds spoken audio in another language. Subtitles display text alongside the video. You can combine both. For accessibility and viewing without sound, subtitles remain useful, even when your video already has a clear Dutch voice-over.

Are lip movements automatically adjusted during dubbing?

Don’t assume that’s the case. ElevenLabs states that dubbing does not include lip-syncing. The translated voice may therefore not align perfectly with visible mouth movements. Be sure to check close-ups of speakers. In explanatory videos with screen shots, this is usually less of an issue than with a presenter who is constantly on camera.

Can you correct the translation later?

Supported dubbing features allow you to edit transcripts and translations and regenerate segments. First, check the genuine transcript: an error there can carry over into the translation. When making changes, take the available speaking time into account so that the brand new segment doesn’t sound rushed.

What does Voice Changer do?

Voice Changer alters the voice in an existing spoken recording while attempting to preserve characteristics of the performance, such as rhythm and expression. This is useful when you want to control the delivery yourself. It is not a substitute for a good source recording; poor intelligibility may still be an issue even after conversion.

What is ElevenLabs Studio?

Studio is a project environment for audio and video productions. Among other things, you can bring together scripts, voice recordings, and various media files. For longer stories or multiple speakers, this offers a better overview than separate folders. Use a clear organization and recognizable file names so that changes can be easily tracked later.

Can ElevenLabs convert spoken audio to text?

Yes, with Speech to Text, you can have recordings transcribed. Check names, technical terms, and passages with multiple speakers or background noise. A transcript is a useful starting point, but it is not a guaranteed verbatim record. For quotes or important instructions, you should compare the text with the recording.

Can you also create sound effects and music?

ElevenLabs offers features for sound effects and music. Check the available options and the specific terms of use for those products. Use background sound sparingly during spoken explanations. The voice-over must remain audible on standard speakers, not just on high-quality headphones.

Free versions, subscriptions, and credits

Is ElevenLabs free?

Yes. The free subscription currently includes 10,000 credits per month and access to a selection of features. It is intended for exploring the platform’s capabilities and is not for commercial use. Be sure to review the key terms before publication; being able to download an audio file does not mean that all uses are permitted.

What ElevenLabs subscriptions are available?

For creative applications, options include Free, Starter, Creator, Pro, Scale, and Business, with Enterprise available for custom solutions. Differences lie in storage space, voice clones, quality, and collaboration features. Be sure to check which products you’re comparing: pricing structures for creative content, agents, and API usage may vary.

How much does ElevenLabs cost?

The regular monthly prices for the Creative subscriptions currently start at $6 for Starter and $22 for Creator. Pro costs $99 per month. Introductory offers may vary; the prices listed do not include taxes. Always check the price after the promotional period and the total at checkout.

What are ElevenLabs credits?

Credits are units of use for the available products. For text-to-speech, usage may depend on factors such as the number of characters and the selected model. Other features are calculated differently. Therefore, do not treat a credit as a fixed number of spoken seconds or as a single full generation.

How many minutes of audio do you get for your credits?

That depends on the model, voice, function, and text. The minutes listed on a pricing page are usually an estimate. Speaking speed and phrasing also influence the final duration. Create a representative test and review the usage before calculating the costs of a large-scale production.

Are unused credits carried over?

For eligible subscriptions, unused credits can be carried over, up to the applicable limit. ElevenLabs allows for a maximum of two months’ worth of unused credits to be carried over. One of the conditions is that your subscription remains active. Downgrading, canceling, and credits purchased separately have different consequences and rules.

What happens when your credits run out?

You can wait for the next renewal, upgrade, or use an available payment option for additional usage. The options you see depend on your account and subscription type. Check whether additional charges may be incurred automatically. This way, you can keep your expenses under control during a larger production run.

What subscription do you need for voice cloning?

Instant Voice Cloning is available starting with the Starter plan. For Professional Voice Cloning, you’ll need the Creator plan or higher. In addition to checking your access level, also verify the number of available voices and your expected usage. Being able to create a clone does not mean you can generate unlimited audio with that voice.

Commercial use and publication

Can you use ElevenLabs audio for commercial purposes for free?

No. The free subscription does not include a commercial key. Therefore, do not use audio generated for free for purposes such as advertisements, paid client projects, or other commercial purposes. If you intend to use the material for business purposes, select an appropriate paid subscription before generating it.

Do you need to credit ElevenLabs for free publication?

Yes. For permitted non-commercial publication of free-to-generate audio, ElevenLabs requires attribution, typically “elevenlabs.io” or “11.ai” in the title. Different rules apply to some products. Check the current publication rules for the feature you used to create the material.

Can you still use the old free audio after you pay?

An upgrade does not automatically convert audio you generated for free into a commercial key. Regenerate the necessary audio while subscribed to an appropriate paid plan. For client work, also save your project data and generation date so you can later verify the terms under which the material was created.

Can you continue using the audio after you cancel your subscription?

According to ElevenLabs, the commercial key for eligible content you’ve created during a paid subscription remains valid, subject to the applicable terms and conditions. Be sure to download your final files yourself. The service does not guarantee that previous content will remain accessible indefinitely if you switch back to the free subscription.

Are beta features suitable for commercial productions?

Not automatically. ElevenLabs excludes commercial and production use of Beta Services in accordance with the applicable terms and conditions. Therefore, please check the status of a feature before using it for paid work. A paid subscription does not remove individual restrictions on a beta feature.

Privacy and Troubleshooting

Does ElevenLabs use your data to improve its models?

ElevenLabs states that certain submitted data may be used to improve audio models. Different standard terms apply to Enterprise. You can adjust your data preferences. Distinguish between processing for your order, storing files, and use for model improvement; these are separate issues.

How do you disable model training?

Open your profile and go to Terms and Privacy > Data Use. Turn off “Improve the models for everyone” and save your selection. According to the documentation, newly submitted material will not be used for model training after that. Do not assume that this change applies retroactively.

Can you upload confidential customer texts?

First, review the customer agreements and terms and conditions for your account. Remove any unnecessary personal data, and use fictitious information for voice tests. A disabled training option does not address every question regarding storage, access, and processing. Evaluate those agreements separately when confidentiality is important.

What should you do if ElevenLabs isn’t working?

Save your text, check your connection, and refresh the page. Then try using a modern browser or an incognito window. Also check if you have credits available and if the voice feature works for the model you selected. Make a note of any error messages before contacting support.

What should you check before publishing a voice-over?

Listen to the final file in its entirety and compare it to your script. Be sure to check names, amounts, and instructions. Also, listen through regular speakers. A recording may sound good on headphones, but it might be too quiet on a phone or get drowned out by background music.

• Content: Are the words, amounts, product names, and steps correct?
• Pronunciation: Are the abbreviations, names, and foreign terms understandable?
• Finishing touches: Are the tempo, pauses, and volume appropriate for the context?
• Rights: Have the subscription, voice usage, and publication terms been verified?

Which ElevenLabs feature is best suited for which task?

Key insights about ElevenLabs

• You can get started without recording your own voice: choose an available voice and test a short, well-written script.
• Dutch is supported: check accents, pronunciation, and suitability using your own text.
• Voice cloning has specific requirements: for Professional Voice Cloning, the voice must be your own and must be verified.
• Business use requires an appropriate key: free audio and beta features have significant limitations.
• Due diligence is still required: review content, audio, privacy settings, and publication rights before distributing anything.