Version 2.6.2 ¡
Vois Studio 2.6.2 keeps your license key when a subscription expires and gives quiet built-in voices the usual level. When you copy a script, the speakers stay with the text.
If your subscription expires, Vois Studio keeps your license key on this computer. When you pay for the subscription again, Vois Studio restores your plan, and you do not enter the key again. If you bought a new subscription, select "Use a different license" and enter the new key. While the subscription is expired, the paid features are not available, also offline.
The built-in Sonic voices and the Fast voices Marius and Javert speak at almost the same level as other voices. Before, they were a maximum of 11 dB less loud. The Sonic voice samples have the new level too. Clips that you made before the update keep their level. To give a clip the new level, delete the clip, then use Generate Audio.
Copied text from the script editor includes the speakers, emotions, pauses and notes. When you paste the text into the script editor again, you get them again.
If Vois Studio cannot use the voice of a speaker, Generate Audio shows one message and opens the voice list. The lines of the other speakers continue. If a speaker has no voice, Vois Studio tells you to assign one. It does not use the voice of a different speaker.
New audio from all engines stays at -1 dBTP or less. Clips and exported files that you made before the update do not change.
If the audio of a line is almost silent, Vois Studio does not keep it [E3010]. The message tells you to make the line longer or to select a different voice.
A backup operates when the data folder of Vois Studio is a link to a different drive. On Windows, a backup to a network folder no longer fails. After a restore, "Restart now" opens Vois Studio again with the restored data.
If you remove a speaker tag and then type it again or use Undo, the speaker has its voice again. A voice change from vois-cli speaker update stays when you close the script.
If Vois Studio cannot save a script, for example because the disk is full, it keeps your text. When you open the script again, you can select "Restore my text" or "Load saved text".
The character count includes only the text that you type, not speaker names, emotions or pauses. The segment and speaker counts do not include notes and code blocks. The "/" menu opens at the cursor and does not add a space after the speaker name.
Export Script keeps a line break that you added with Shift-Enter. It no longer adds a space after an emotion or pause tag at each export.
On macOS, Command-Q operates when the window is busy and shows the "Stop Vois Studio?" dialog. When you close the window, a click on Vois Studio in the Dock opens it again. The app menu and the About window show the name "Vois Studio". On all computers, you can drag the toolbar to move the window.
When your computer does not have sufficient memory for Expressive, batch generation stops at the first line that it cannot generate. It shows one message. On Windows computers with an NVIDIA graphics card, Expressive stops less frequently. If Expressive stops on the graphics card, the next start uses the processor and shows a message.
In "Settings > License", Manage Billing and Upgrade to Pro open the Polar customer portal. These buttons show only for a subscription that you bought through Polar.
If a save stops, no part of the file stays at the name that you selected. After you select a file or a folder in a dialog, Vois Studio closes the folder when the operation ends. Then you can eject the disk. On macOS, an exported audio file gets the permissions of a new file in that folder.
An audio file that you add to the timeline shows its file name. A sound from the audio library shows its name, for example "Whoosh".
After vois-cli script generate, the commands vois-cli script segments and vois-cli export compile-script operate. The command vois-cli script generate uses the engine that you selected in the app. The command vois-cli model status shows all engines that are in memory. When audio is almost silent, a command stops with exit code 5.
The speaker panel no longer shows an incorrect note for the built-in Fast voices in Italian, German, Spanish and Portuguese. In Listen > URL, Clear removes the "Extraction Failed" message. Vois Studio starts when you open it through a link to its folder. On Windows, the uninstaller removes all the files that it installed.
After the update, vois-cli script generate makes the audio of unlocked lines with the built-in Sonic voices, Marius or Javert again. The new audio has the new voice level.
Listen does not open a page on your computer or on a local network. For example, it does not open localhost or printer.local. It stops at a page that is larger than 5 MiB. For these pages, Listen does not use the proxy of your computer.
Version 2.6.1 ¡
Vois Studio 2.6.1 gives quiet voices the usual level and removes noise from mastered exports. Exports do not stop after a day offline, and locked lines stay locked.
Some library voices, premium voices and cloned voices were very quiet. They speak at almost the same level as other voices. This applies to all engines, but not to Sonic. Clips that you made before the update keep their level. To give a clip the new level, delete the clip, then use Generate Audio.
A Fast line in German, Spanish, Italian or Portuguese can use a language that is not on your computer. Vois Studio downloads that language first, then generates the line.
When vois-cli changes a script that is open, the app shows the new text. If you have edits that you did not save, you select the text to keep.
After more than a day offline, exports and vois-cli commands no longer stop and wait for the plan check.
If the licence server has a fault, your paid plan continues for the offline period.
A mastered export no longer adds high-frequency noise to the voice.
If you change the text of a locked line, the padlock stays. You can unlock the line. Lock line and Takes change only the line that you selected.
If Vois Studio cannot lock a clip, the message tells you why. When Generate Audio does not change a locked line, a message tells you one time in each run.
The audio library shows the licence of its music and sound effects, the Pixabay Content License.
vois-cli finds a file that you give with a short path, for example source.wav.
After the update, vois-cli script generate makes the audio of all unlocked lines again. The new audio has the new voice level.
Version 2.6.0 ¡
Vois Studio 2.6.0 brings voices you buy into your Library sooner, explains clearly when it canât start, and makes Deactivate and backup restore do exactly what they say.
Voices bought on another computer or through a partner offer now show up the next time you open Vois Studio, or within a few minutes when you switch back to it.
If voices you bought are missing from your Library, Voices now asks you once to confirm your email. Vois Studio sends a 6-digit code to the email on your licence, and your voices appear as soon as you enter it.
When several speakers share a voice that canât be used, Generate Audio shows one message for that voice, saying how many speakers it affects, instead of one message per speaker. The other speakersâ lines keep generating.
If Vois Studio canât finish starting, it now says so and what to try next, instead of closing without a word.
Deactivating your licence in Settings now works even when this computer was already released, instead of showing an error and keeping the licence here.
Restoring a backup now replaces this computerâs generated audio too, even when the backup has none, and a backup made with Vois Studio 2.4 or earlier no longer leaves that versionâs old voice files behind. Your recordings are kept.
Exporting from the command line with loudness normalization now uses the loudness you saved in your mastering settings, and every export says which loudness it used.
Version 2.5.3 ¡
Vois Studio 2.5.3 saves the complete recording when you export a whole audiobook, keeps long scripts generating at full speed, and lets Cancel stop an export at any stage. The counts beside Generate Audio now match each run.
Generating a long script, such as a whole book, no longer slows down as lines finish.
Exporting a long audiobook with mastering on now uses far less memory while checking the finished fileâs loudness.
Exporting a very long project, such as a whole audiobook, now saves the complete recording instead of stopping partway through without an error. A WAV export larger than 4 GB is saved as RF64, the standard format for WAV files of that size.
Cancel now stops an export within about a second, whether it is reading audio, adding the watermark, checking or encoding the finished file, and nothing is saved.
Long built-in library tracks such as âAmbient Backgroundâ now export. When an audio file is there but canât be played, the timeline and the export error say âCanât read fileâ instead of âFile not foundâ.
Generate Audio now tries that speakerâs lines once and says once which speaker and what to do, while the other speakersâ lines keep generating. A blank notification no longer appears while failing lines are reported.
The counts beside Generate Audio now match the run: âcompletedâ counts only lines that got new audio, and âfailedâ lists every failed line instead of stopping at 50. Pressing Generate Audio when every line is up to date now says so.
When a voice engineâs downloaded files are damaged, Vois Studio now says so and Download in Voice Settings repairs it.
Checking a file for a Vois watermark now works on WAV exports larger than 4 GB and reads the whole of a recording longer than about three hours.
Reopening an AppSumo activation link that was already used now explains that the link works once and how to get a fresh one from AppSumo.
Version 2.5.2 ¡
Vois Studio 2.5.2 lets you switch a line back to an earlier take and lock the lines you like, so Generate Audio keeps them. Timeline waveforms show real speech and pauses, activating on a new computer explains each step, and backups move between Mac and Windows.
Liked an earlier version of a line better? Right-click its clip and choose Takes to switch back to one of its earlier takes. Picking one locks the line, and Undo puts everything back.
Lock a line from the clip menu, or press L, and Generate Audio leaves its audio alone. Locked lines show a padlock on the clip and in the script.
Timeline clips show their real waveform, drawn calmly in the lower half of each clip, so pauses and cut-off endings are visible at a glance.
When Vois Studio needs you to confirm a new computer by email, it now shows which inbox to check, lets you resend the email and finishes by itself once you confirm, instead of showing an error. Try it Free with an email that already has a licence now sends a code that signs you in.
A backup made on a Mac now restores on Windows, and a backup made on Windows restores on a Mac.
System cleanup apps can no longer remove the audio in your projects, and imported audio in older projects is no longer marked missing while it still plays.
A script of a sentence or two in English is now recognised with Auto language, instead of asking you to choose a language first.
Creating a project past the Free limit now explains the limit and offers Upgrade instead of doing nothing. A paid plan that canât be confirmed is asked to connect and retry, not to upgrade.
After you click a clip, its shortcuts (P, S, T, L and Delete) act on the clip instead of typing into the script, and selecting a clip no longer adds an empty Undo step.
Starting vois speak again after an earlier attempt found no audio output device no longer closes Vois Studio.
Version 2.5.1 ¡
Vois Studio 2.5.1 highlights each word as it plays, saves exports at the quality and loudness you choose, and keeps your timeline edits when you generate audio again. Lines that stall partway are made again instead of left silent, and restoring a backup works on Windows.
As audio plays, the script highlights each word as it is spoken, in the editor and in Listen.
WAV and FLAC are saved at the bit depth you pick (16-bit, 24-bit or 32-bit float), and MP3 and AAC at the bitrate you pick. Before, WAV and FLAC were always saved as 16-bit and MP3 came out at about 77 kbps. The export window only offers combinations each format can save, and tells you when changing format moves a setting.
Exports with loudness normalisation now reach the loudness you choose, within 0.3 LU, and stay under the peak limit. Before, they came out 2.5 to 3.3 LU quieter. If a recording canât reach the target without going over the limit, the export tells you.
ACX Audiobook, Apple Podcasts, Spotify Podcasts, YouTube Video and Generic High Quality now save stereo audio at each platformâs own sample rate and bitrate.
In our tests, an 11-minute episode at 96 kHz now exports in about 20 seconds instead of nearly 2 minutes. The progress bar keeps moving while the audio is mastered, and Cancel stops an export that is in progress.
When an export has a problem, the warning names it and the line it is on, with a button to go to that line, and Export anyway takes one press. A long export no longer leaves a false âExport never completedâ warning behind.
If audio your project uses is missing, the export stops and names what to fix instead of saving a file with silence in it. Lines with missing audio are marked on the timeline, and Generate Audio makes them again. Music, sound effects and imported audio whose file was moved or deleted are marked âFile not foundâ with the fileâs name, so you can put it back or remove the clip.
After you change a voice or a line, the clips that need new audio are marked, and the Generate Audio count matches them. During a run, clips show when they are queued and generating. If you stop, clips not yet done keep their previous audio and stay marked.
Generate Audio keeps trims, splits, fades and volume on lines that didnât change, and lines you removed stay removed. A changed line plays its full new take, and the clips after it keep their spacing without overlapping.
The failed count next to Generate Audio opens a list of the lines that failed, each with its reason, and takes you to a line when you choose it. The counts start at zero on each run.
Style directions are offered only on lines that use Craft or Omni. A direction on a line spoken by another engine stays in your script, shown muted, and works again when that line uses Craft or Omni.
Clicking a timeline clip takes you to its own line in the script. Clip waveforms show again. Undo in the timeline no longer undoes script edits, and renaming a script is kept after you restart.
When a lineâs audio stalls partway or is cut short, Vois Studio makes that part again with the same voice and engine. If it still canât, the line is marked as failed with the reason, instead of keeping a long silence or missing words.
Craft voices no longer turn warbly or unsteady partway through a line, and sound closer to the voice you chose. Lines you already generated keep their audio; generate a line again to hear the difference.
Text that ends in a quotation or contains an ellipsis is read as written, without a stray sound, an extra word such as âdotâ or an extra pause.
When Vois Studio needs to confirm your paid plan before an export, it checks online first. If it canât, it asks you to connect to the internet and retry, instead of offering export credits.
The CLI no longer says CLI access is disabled on a plan that includes it, such as AppSumo Lifetime Tier 3, after Vois Studio has gone a day without checking your plan. It checks your plan first and, if it canât, asks you to connect and retry.
Stopping a generation in Vois Studio no longer stops agent speech or generations started from the CLI, and Ctrl-C in one CLI command no longer cancels another. vois tts cancel still stops every CLI generation.
vois speak --follow --play accepts text that starts with a short greeting such as âHi Sam.â without --language.
Signing up for Free asks for your name as well as your email.
Restoring a backup on Windows now works. Before, it stopped with a message about writing to a folder. If Vois Studio canât open a backup file, the message now says how to fix it, such as moving the file to a folder on this PC.
Quitting with Cmd+Q, the app menu, the Dock or a logout no longer crashes after you have generated audio.
On Windows computers whose processor lacks AVX2, Vois Studio explains that the voice engines canât run on it instead of closing when you generate.
Vois Studio no longer removes audio your projects still use when it frees up disk space.
Background quality checks no longer interrupt an export with a false âtimeline source changedâ error.
Tutorials now open in your browser on YouTube instead of showing a video player error inside Vois Studio.
Version 2.5.0 ¡
Vois Studio 2.5.0 lets your AI agents read aloud to you, adds Sonic, a fast engine on every plan, and makes Free never expire. Fast now speaks German, Italian, Portuguese and Spanish, Craft on Pro directs any voice with a recording, Omni takes direction in plain words, Voice Design starts from a description, and long Listen items now export in full.
With vois speak, AI agents such as Claude Code, Codex or your own scripts read their text aloud on your computer, phrase by phrase, with a Sonic voice. It works on every plan, Free included, with no daily limit. vois speak only plays: nothing is saved. To save Sonic audio as a file, use vois tts generate --engine sonic, which follows your planâs usual rules.
Sonic speaks 31 languages with its ten voices, on every plan, Free included. Choose it in projects, Listen and the CLI. Library, cloned and designed voices are not available with Sonic yet. Sonic needs at least 4 GB of free memory on Windows and 2 GB on macOS, and Vois Studio tells you when there is not enough.
Fast now speaks German, Italian, Portuguese and Spanish as well as English. Choose the language in the editor, in Listen or in the CLI. The first time you choose a new language, Vois Studio asks you to download it once. Cloned voices keep the accent of their recording.
Craft is a new engine on Pro that follows a delivery direction in plain words, like Omni, on every voice that has a recording: library voices, your cloned voices with a transcript and your designed voices. It speaks Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian and Spanish, and cloned voices keep the accent of their recording. In the CLI, use --engine craft with --instruct. Free can try Craft in the app: generate and listen; exporting it uses purchased export credits. Craft needs about 9 GB of memory on a Mac and 11 GB on Windows.
Headless mode runs Vois Studio on your own Mac or Windows PC with no window open, for agents, scripts, remote shells and scheduled jobs: vois service start, stop and status. Manage your licence from a terminal with vois license activate, status and deactivate. Your planâs usual rules apply.
vois speak playback is open to every plan, including AppSumo Lifetime Tier 1, which does not include the rest of the CLI.
Free has no trial period and no end date. Verify your email to start; no card needed. It includes Sonic and its ten voices, 10 generations a day, and 10 free exports a month of English Sonic audio, which reset on the 1st (UTC). You can also try Fast, Expressive, Multilingual, Omni and Craft in the app: generate and listen. Settings â License shows how many free exports you have left this month.
Purchased export credits export audio from any engine and in any language, and unlock the CLI on a pay-as-you-go basis. Credits are separate from the 10 monthly free exports and do not add commercial rights.
On Free, the export window tells you whether an export uses one of your free exports or a credit before you choose where to save it. If it cannot be exported, it says why and offers Buy credits or Upgrade.
Free is for hobby, research and non-commercial use. Commercial use needs Subscriber or Pro. Free does not include voice cloning or Voice Design.
If the verified email on your Pro account also holds an active AppSumo Tier 2 or Tier 3 licence, your AppSumo-exclusive voices are available on Pro too.
Describe how a line should sound, such as slower and warmer or brighter and more urgent, instead of choosing from bracketed tags like [laughter]. Tags already in your scripts are left as you wrote them. Omni is part of Pro.
Arabic, Burmese, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Khmer, Korean, Lao, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Tagalog, Thai, Turkish and Vietnamese. That is fewer languages than before, and Omni now tells you when it does not speak a language instead of trying.
Describe a voice in your own words, or start from an example and edit it. Voice Design follows who is speaking and how: gender, age, pitch, voice texture, tone, emotion and pace. It cannot set an accent reliably. It designs in 30 languages at one quality, so there is no quality setting to choose. Voice Design is part of Pro.
Start Voice Design from an existing voice to change its pace, emotion and energy while it stays the same speaker.
Designed voices made only from a description with the previous Omni cannot be reproduced by the upgraded Omni. Vois Studio marks them and walks you through re-creating them from their description. Designed voices with a saved recording keep working.
The language rules from 2.4.1 now cover every engine, and a chosen engine always runs as chosen. Omni, Sonic and Craft tell you when they do not speak a language instead of switching engines. Fast asks you to choose the language when it is unsure or does not speak it, instead of reading the text as English, and in the CLI Auto asks you to choose when it cannot tell the language of a short or mixed text.
Vois Studio uses less memory after launch, so the engine you choose is less likely to run short of memory on 16 GB Macs.
A cloned voiceâs card now asks for the words spoken in its recording, and you can edit them on the card. A transcript helps Omni speak the voice better, and Craft needs one to speak it.
CLI output names engines fast, expressive, multilingual, omni, sonic and craft; earlier names still work as input. vois voice clone now needs --transcript, the exact words spoken in the recording. vois model download waits and shows progress. New exit codes report a full Sonic queue, not enough memory, a failed download and a cancelled job. Some model status and model list fields are renamed and vois model variant is removed; the CLI guide has the full list.
Cloned voices keep the language and accent of the recording used to create them, on every engine.
Exporting a long Listen item now saves the whole recording. Before, it saved only about the first 10 seconds, in both MP3 and WAV.
Voice Design no longer fails the first time you use it. The first preview gets it ready and shows progress.
Generating a Voice Design preview again with the same description now gives a new take to compare, instead of repeating the last one.
A voice you save from Voice Design keeps the speaker you chose in the preview, including when you re-export its reference recording. The âWhisperingâ and âWhisperâ examples are now âSoft-spokenâ and âSoftâ, which matches how they sound.
English clock times are read as âthree thirty PMâ, and phone numbers digit by digit, on every engine. Other languages are unchanged.
Long cloned, library and designed voices no longer fail or run out of memory with Fast on Windows. Fast now learns a voice from up to the first 20 seconds of its recording, ending on a pause. Your full recording is kept, every other engine still uses all of it, and voices Fast has already prepared keep their current sound.
Fast now honours --seed like the other engines: the same text, voice and seed give the same audio. The first time you generate a line again on Fast, you get a new take.
Exports shorter than 5 seconds explain why they cannot be exported instead of showing a technical error, and no free export or credit is used. Other export messages, including Free plan limits, are shown in full.
Settings â License names your plan, such as âLifetime Tier 2â, âPro Lifetimeâ or âFounder Proâ, and plans that never renew no longer show a Manage Billing button. A one-year plan bought through DealFuel shows as a one-year plan with the date your access ends.
On a plan without the CLI, such as AppSumo Lifetime Tier 1, vois status says so and points to an upgrade instead of suggesting export credits. Paid plans no longer see âExport Credits: 0â.
Regenerating a line after an export no longer makes the quality check report that line as missing. The editorâs progress counter no longer counts cancelled or failed lines as completed.
When you generate a script again and a lineâs new take fails, the line keeps its earlier take. Before, it lost the earlier audio.
MP3 and AAC exports could stall on some Macs and fail after about five minutes. They now finish normally.
If your connection drops for a moment while Vois Studio downloads a voice engine, the download now picks up where it left off instead of failing during setup. If our servers are busy, Vois Studio tells you so instead of blaming your network.
The export window no longer says a check was âTurned off in settingsâ when you never turned it off. Skipped checks now have clear names and the real reason.
Searching the speaker voice picker for a multi-word name, such as âMy Designed Voiceâ, now finds cloned and designed voices.
The timeline and Listen player no longer show a position past the end, such as â0:09 / 0:08â.
Version 2.4.1 ¡
Vois 2.4.1 is a fix release. Generation on Windows is more stable, short English lines stay in English, and Vois never swaps the engine you chose.
Vois now checks your graphics card before using it for voices. If the card has little memory of its own, or its NVIDIA driver is too old, Vois uses the processor instead and tells you once. If the graphics card ever stops the app, the next launch uses the processor automatically.
Headings and single sentences no longer switch to another language because they look like German or Italian. Short lines follow the language of the whole script.
If Fast or Expressive is selected and your text is clearly in another language, Vois still uses your choice and suggests Multilingual once.
In Auto mode, if a clip needs an engine that is not available, only that clip fails. The rest of the batch keeps going.
Long sentences are split evenly, so one or two words are no longer left to be spoken on their own after a pause.
Lifetime plans that need their licence revalidated now say so, instead of asking you to buy credits. Stopping a generation is no longer reported as an error.
Version 2.4.0 ¡
Your built-in voices already spoke multiple languages, but an English accent could carry into other languages. Vois 2.4.0 improves accent, rhythm and style across the 23 supported Multilingual languages, helping your chosen speaker sound more natural in your scriptâs language. You also get six new speakers, easier voice selection and fixes that keep your projects moving.
Ellis, Mara, Caleb, Nora, Jules and Reed join the cast, bringing the core library to 112 voices. More choice for your next narrator, podcast guest or character.
The improvements to accent and style also apply to the 100 AppSumo voices if they are included in your licence. Your existing access stays the same.
Audition your chosen speaker in the language of your project before generating a full passage. Language filters in the Voice Library, Favorites and website voice browser make it easier to compare voices.
These improvements apply to built-in voices and the AppSumo collection included with your licence. Cloned voices retain the language and accent characteristics of the recording used to create them. For the most natural delivery, use a voice cloned in the language of your script.
Fewer wrong words in affected Mandarin sentences, more accurate Arabic word forms and improved pronunciation in Greek, Turkish and Korean. Several Japanese readings are corrected too.
Longer passages are less likely to stop early, repeat phrases or stray from your script. Narration in German, French, Hindi, Italian and Japanese follows the complete text more reliably, with fewer long, unwanted pauses.
Numbers, prices and abbreviations are spoken more consistently in the selected language instead of unexpectedly switching to English. Arabic narration with Omni is also more reliable.
Louder Omni passages are less prone to distortion. Fixes to WAV export and copy preserve the quality of your generated audio.
These corrections are not a promise of native-speaker pronunciation. Individual words can still be read unexpectedly, including Chinese words with multiple readings, Hebrew without vowel points and some Japanese text. Listen to generated speech before publishing.
Listen can choose the language automatically. Use the manual override when you need a particular language, especially for short or ambiguous text.
Choose the language beside Save in the editor. When working through the CLI, let Vois detect it or set it yourself for short passages, mixed-language scripts or romanized text.
Switch projects or leave the editor without interrupting speech started through the CLI. Previews also work more reliably for saved cloned voices.
Filter by male or female voice alongside language, category and search in the Library, Favorites and website voice browser. Favorites now follow your filters correctly, and included AppSumo voices are easier to identify.
Each built-in voice now has its own personalized avatar, with male and female voices represented correctly. The same avatar appears in the app and on the website, so your chosen speaker is easier to recognize wherever you browse.
Choose an avatar when cloning a voice or saving a voice made with Voice Design. Your choice follows that voice throughout the app. You can also set and edit its language and gender labels. Cloning and Voice Design remain subject to your existing plan access.
A more compact Library toolbar brings filters and sorting together in card and list views. Audiobooks and Podcasts are easier to find, and the editorâs language choice sits beside Save.
Select and delete multiple cloned or designed voices together, including a filtered selection. Review your choice before deleting, and retry any voices that could not be removed.
Fixed the legibility issue that made the selected model quality hard to see. Efficient and High quality now have a clearly visible selection in both light and dark themes, with keyboard support.
Clearer messages explain when a voice download has failed or the selected language is unsupported, so you know whether to retry or choose another language.
Fixed generation failures after computer cleanup and improved voice-cloning reliability when optional notes are left blank.
Licence refreshes no longer incorrectly sign you out or interrupt access to your voices.
Explore audiobook, podcast and voiceover workflows on the rebuilt homepage, watch the demo, and find voices more easily with language and gender filters.
Tutorial cards now open the video you selected. Updated guides help you choose languages, direct speech and create multilingual projects through the app or CLI.
These desktop improvements were already included in the previous release notes. They remain part of Vois, but are not new in 2.4.0.
The Windows installer and app identify SEED BLOCKS PTY LTD as the verified publisher, so you can check who supplied your download.
More accurate readings for Japanese text containing kanji.
Web imports no longer appear to fail when they finish quickly. After importing, you return to the editor correctly, and saved Listen items keep their original source details.
Switching between projects or scripts no longer shows content from an earlier selection.
Music and sound effects fit the narration when first added, then keep their placement and duration when you edit or regenerate narration.
Waveforms remain usable on long timelines. Playback no longer jumps to an earlier position or restarts after you stop it when controls are used in quick succession.
If your computer does not have enough available memory to begin, Vois stops the job with an explanation instead of repeating the same error for each remaining clip.
If an audio file changes while you are exporting, Vois explains that you need to retry.
Retrying an interrupted export no longer risks deducting the same credits twice.
Longer multilingual generation jobs can finish without the CLI ending its wait too early. This improves reliability, not generation speed.
These download and onboarding changes also predate 2.4.0.
Apple Silicon Mac users are more reliably offered the correct download, including when using privacy-focused browser settings.
Existing activation links continue to provide the licence key and current download instead of leading to a dead end.
Update Vois Studio on your Apple Silicon Mac or Windows PC. Stay online while Vois Studio prepares your voices for their first use. Listen to a preview of the speaker that you select, in the language of your script. Listen to the completed audio before you publish it.