AI voiceover generation lets you add professional narration to your topics without recording anything yourself.
You write a script, choose a voice, and Tribal Habits generates the audio for you. There is nothing to upload, edit or re-record, and you can change a script at any time and generate fresh audio in seconds.
You can add AI generated audio to the following elements:
In Narration, Interact and Hotspot elements you set a script and voice for each slide, item or hotspot, so one element can contain several audio files. All other elements use a single script and voice.
Tribal Habits uses generative AI voices for most languages. These voices are noticeably more natural than the previous generation, with better pacing, emphasis and intonation, which makes longer scripts much easier to listen to. See the Available voices section below for more details.
Note: AI voiceover files are generated and stored on Australian-based servers.
Key concepts
Voice quality: Voices are available in three levels of quality. Generative voices are the newest and most natural sounding. Neural voices are the previous generation and still sound clear and professional. Standard voices are the oldest and are only used where no newer option exists for that language. You do not select a quality level. It is set automatically by the voice you choose.
Regenerate: Creating a brand new audio file from the same script and the same voice. Generative voices produce a slightly different take each time they run, so regenerating gives you a fresh version without having to change a single word of your script.
Steps to add AI voiceover to an element
Open your topic in the creator and select the element you want to add audio to.
Display the Voiceover options by ticking the Add audio to item box in most elements.
Note that the Voiceover fields are visible by default in the Narration and Audio elements.
Choose Automatic to generate the audio using an AI voice, rather than uploading or recording your own.
Select a voice from the Character list. Each voice displays its name, language code and gender, for example 'Olivia (en-AU, Female)'.
Enter your script in the text box. Write the script the way you want it spoken, including punctuation, as this shapes the pacing of the audio.
Select Play audio to generate and hear a preview. The first play takes a few moments while the audio is created. After that, playback is immediate.
If any part of the audio does not sound right, select Regenerate next to the preview audio player to create a new version, then listen again.
Save the element.
Important: Generative voices produce a slightly different take each time, and very occasionally a take includes a small defect such as a clipped word or an unnatural pause. This affects a small number of files, but it is the reason the Regenerate option exists. Always listen to your audio in full before publishing, and regenerate anything that does not sound right.
Available voices
Voices are grouped below by quality level. The language and gender of each voice is fixed and cannot be changed.
Generative voices
Olivia en-AU Female
Aria en-NZ Female
Amy en-GB Female
Brian en-GB Male
Niamh en-IE Female
Ayanda en-ZA Female
Jasmine en-SG Female
Kajal en-IN / hi-IN Female
Danielle en-US Female
Joanna en-US Female
Ruth en-US Female
Salli en-US Female
Matthew en-US Male
Stephen en-US Male
Lupe es-US Female
Pedro es-US Male
Lucia es-ES Female
Sergio es-ES Male
Mia es-MX Female
Andres es-MX Male
Lea fr-FR Female
Remi fr-FR Male
Vicki de-DE Female
Daniel de-DE Male
Bianca it-IT Female
Camila pt-BR Female
Seoyeon ko-KR Female
Neural voices
Emma en-GB Female
Arthur en-GB Male
Kendra en-US Female
Kimberly en-US Female
Gregory en-US Male
Hala arb Female
Zayd ar-AE Male
Zhiyu cmn-CN Female
Hiujin yue-CN Female
Kazuha ja-JP Female
Takumi ja-JP Male
Jihye ko-KR Female
Ines pt-PT Female
Thiago pt-BR Male
Adriano it-IT Male
Burcu tr-TR Female
Standard voices
Tatyana ru-RU Female
Best practices
Write your script the way you want it spoken. Short sentences, commas, and full stops all help the voice pace itself naturally.
Listen to every script in full before publishing. This is the single most effective check you can make.
Check how numbers, dates, acronyms and unusual names are pronounced. If an acronym is read as a word rather than letters, try spacing or punctuating it differently, for example 'W H S' instead of 'WHS'.
Regenerate rather than rewrite. If you are happy with your script but not the take, use Regenerate first.
Use one voice consistently across a topic unless you are deliberately using different voices for different characters or perspectives.
Preview your audio again after editing a script, as a new file is generated whenever the script or voice changes.
Troubleshooting
Problem: A word is clipped, rushed or mispronounced, or there is an odd pause in the audio.
Cause: Generative voices create a new take each time they run, and occasionally a take contains a small defect.
Solution: Select Regenerate and listen again. If the same issue occurs repeatedly in the same place, adjust the punctuation or wording of that sentence in your script.
Problem: The voice you want is not available in a particular language.
Cause: Not every language has a generative voice available. Some languages use neural or standard voices instead.
Solution: Check the voice lists above and select the closest available voice for that language.
Problem: Nothing plays when you select Play audio.
Cause: The element has no script entered, or no voice selected.
Solution: Confirm both the script and the voice are set, then try again. The first play of a new script takes a few moments while the audio is generated.
FAQs
Do I need to save or publish a topic after regenerating audio?
No. The new audio file is created and stored as soon as you regenerate, and learners will receive the newest version.
Will my existing topics change to the new generative voices automatically?
No. Audio that has already been generated stays exactly as it is, so published topics will sound the same as they do today. New audio uses the new voices, and you can select Regenerate on any existing element to update it.
Why does the same script sound slightly different each time I regenerate?
Generative voices are not fixed recordings. Each take is produced fresh, so pacing and emphasis vary slightly. This is what allows regeneration to resolve an unwanted take.
Can I use more than one voice in a single element?
Yes, in Narration, Interact and Hotspot elements, because each slide or hotspot has its own script and voice.
Can I upload my own audio instead of using an AI voice?
Yes. Every element that supports audio also lets you upload an .mp3 or .wav file up to 100MB, or record your own audio directly in the creator.
Is there a limit to how long a script can be?
Scripts are best kept to a manageable length per slide or element. Break longer content across multiple slides so learners can follow along and so you can regenerate a single section if needed.
Where is the audio generated and stored?
All AI voice audio is generated and stored on Australian-based servers.