Novidades

Voice Delta Changelog

Todas as atualizações do estúdio web e do aplicativo de desktop para Windows — em linguagem simples, sem jargão técnico.

Aplicativo de Desktopv1.9.0September 19, 2026

Desktop App 1.9.0 — Hindi Cloning and Voice Picker Improvements

New: Hindi voice cloning is available in the Windows app, with Hindi preview text and adjustable pauses between lines.
Improved: Edge voices now show their full voice name and language or region, and can be searched by either.

Webv1.9.0September 19, 2026

Website — Hindi Voice Cloning

New: Clone voices in Hindi from the web studio, preview the clone with Hindi text, and adjust the pause between Hindi lines.

Aplicativo de Desktopv1.8.0August 30, 2026

Desktop App — Multi-Language Voice Cloning

New: Clone voices in 6 languages — English, French, German, Portuguese, Italian, and Spanish. Pick a language when you clone a voice, and generate audio in that language from it.
New: Language pack download manager — the first time you use a non-English language, the Clone tab shows a one-time download for that language's voice model, with a live progress bar. Pause and resume anytime without losing progress.
Fixed: long scripts no longer fail with "Generation timed out". Each generation used to get a fixed time budget based on an estimate that assumed a fast PC and undercounted how many pieces your script splits into — so a long script on an average computer was cut off part-way even though it was generating perfectly well. Generations are now judged on whether they're still producing audio, not on how long they've been running: as long as it keeps making progress, it runs to the end however long that takes.
Fixed: one timeout no longer caused the next. A generation the app had given up on kept running in the background using your whole CPU, so pressing Try again started a second one on an already-busy PC — which is why timeouts tended to come in a run rather than alone. Abandoned generations now stop and hand the CPU back.
Improved: when a generation genuinely does stall, the error now tells you how long the voice engine was silent and how far the job got, instead of just a time limit.
Improved: progress bars start at the real number of pieces your script was split into, so they no longer jump or stall near the end on scripts with lots of paragraphs.
Fixed: deleting a generation while it's still running now actually stops it, instead of leaving it to finish in the background.

WebAugust 30, 2026

Website — Multi-Language Voice Cloning

New: Clone voices in 6 languages — English, French, German, Portuguese, Italian, and Spanish. Pick a language when you clone a voice, and generate audio in that language from it.

Aplicativo de Desktopv1.6.0August 20, 2026

Desktop App — Bulk Generator, Folders & Smarter Captions

New: Bulk Text-to-Voice Generator — add multiple .txt files or a whole folder at once, set one voice (and speed/pitch) for everything or customize each file individually, then generate them all automatically, one at a time.
New: Auto-save your bulk output — point Bulk Generate at a folder once and every finished file saves there automatically, named to match its source .txt file exactly.
New: Retry failed files individually or all at once — and if you hit your daily character limit mid-batch, the queue stops cleanly instead of every remaining file failing one by one.
New: Organize generations into folders — create a folder before generating, or move any past audio into one from History.
New: Download every audio in a folder, or a hand-picked selection, as a single zip.
New: Rename any generated audio at any time — the name sticks even after you generate more audio, it no longer gets overwritten.
Improved: Captions (SRT/VTT) are now timed against the real pauses in your audio instead of an even character-based guess, so subtitles stay in sync on longer scripts.
New: Login sessions are now limited per plan (e.g. standard plans allow 1 web + 1 desktop session at once) — logging in elsewhere signs the older session out automatically with a clear message.

WebAugust 20, 2026

Website — Folders, Download All & Smarter Captions

New: Save your generated audio into folders — create a folder before generating, or move audio into one afterward from History.
New: Download every audio in a folder, or a hand-picked selection, as a single zip — no more one-by-one downloads.
New: Rename any generated audio at any time — the name sticks even after you generate more audio.
Improved: Captions (SRT/VTT) are now timed against the real pauses in your audio instead of an even character-based guess.
New: Login sessions are now limited per plan (e.g. Web plans allow 1 web + 1 desktop session at once; Desktop-only plans allow a single device) — logging in elsewhere signs the older session out automatically.

Aplicativo de Desktopv1.5.0August 13, 2026

Desktop App — Script Markup, Pronunciation Lexicon & Captions

New: Pause Control — insert exact pauses in your script with simple markers (weak, clause, sentence, paragraph, or a custom duration like 500ms).
New: Say-As — control how numbers, dates, times, phone numbers, and units are read out, e.g. write a date and have it spoken naturally.
New: Emphasis — wrap any word in asterisks to have it spoken with natural stress (Natural Flow and Cloud voices).
New: Heading detection — lines starting with # get a natural pause before them, great for structuring long scripts.
New: Pronunciation Lexicon — teach the app how to say specific words or brand names correctly, every time, across every script.
New: Download captions — export SRT or VTT subtitle files alongside your generated audio, ready for video editors.
New: Name your own files — set a custom output name instead of the generic 'voicedelta-output', applied to the audio file and its captions.
Improved: Speed slider now goes up to 3x (previously 2x) and is a smooth slider instead of fixed buttons.
Fixed: accounts were being randomly signed out during active use — sessions now stay alive for as long as you're using the app.
Fixed: high CPU usage on Natural Flow voices that could slow down the whole PC during generation.
A new in-app Docs page explains all the new script markup features with examples.

WebAugust 13, 2026

Website — Script Markup, Pronunciation Lexicon & Captions

New: Pause Control, Say-As, Emphasis, and Heading detection — the same script markup toolbar as the desktop app, now in the browser.
New: Pronunciation Lexicon page — build a dictionary of custom pronunciations that applies automatically to everything you generate.
New: Download SRT or VTT captions for any finished audio, directly from the Studio.
New: Custom output file names for your downloads, instead of a generic filename.
Improved: Speed slider now goes up to 3x.
Fixed: an issue causing some accounts to be randomly logged out while actively using the site.

Aplicativo de Desktopv1.4.0July 29, 2026

Desktop App — Generation Reliability Fixes

Fixed the progress bar getting stuck at 0% on Natural Flow voices — it now updates live as each part of your script is generated.
Generation no longer hangs forever if something goes wrong — it now fails with a clear message after a few minutes so you can simply retry, instead of watching a frozen progress bar.
Fixed voice cloning taking a very long time (or appearing stuck) with long reference clips — clips over 60 seconds are no longer accepted, since only the first 30 seconds are ever used anyway.
The installer now checks for a required system component and installs it automatically if missing, for more consistent performance across different PCs.

Aplicativo de Desktopv1.3.0July 13, 2026

Desktop App — Reliability Update

Fixed a crash on long voice generations that could happen on PCs with less RAM — you can now generate longer scripts and podcasts reliably.
Fixed generation stopping if your laptop's screen turned off or the computer went to sleep mid-generation — the app now keeps working in the background and picks back up automatically.
More reliable generation — if a voice line hits a temporary glitch, the app now tries harder to recover it before giving up.

Aplicativo de Desktopv1.2.0July 7, 2026

Desktop App — Big Quality & Convenience Update

Preview any voice with one click — every voice in the list now has a play button that speaks a short sample before you choose.
See your subscription in the app — Settings now shows your plan, purchase date, and expiry date.
Contact support directly — new WhatsApp button in Settings.
Plan changes apply instantly — upgrade or renew on the website and the app updates within seconds. No more logging out and back in.
Complete audio on long scripts — fixed an issue where words or paragraphs could be missing from very long generations.
Cleaner sound — removed tiny clicks and breaks between sentences in Studio voices and cloned voices.
Fixed the Voice Library page showing up empty.
More reliable generation — temporary network hiccups no longer fail your audio; the app retries automatically.

WebJuly 7, 2026

Website — Desktop Plans & Easier Checkout

New Desktop App Exclusive plans — featured at the top of the pricing page with a direct download button.
Buy with Gumroad without creating an account first — checkout in seconds, your plan activates on the email you use.
Redesigned pricing page with a clearer, centered layout.
New Web vs Desktop comparison on the homepage so you can pick what fits you best.
This changelog page — we'll post every user-facing update here.

Aplicativo de Desktopv1.0June 2026

Desktop App — Public Release

Voice Delta Desktop Studio for Windows — the full studio running on your own PC.
TTS Studio, Podcast Studio, Voice Cloning, and Voice Library — all included.
Generate offline with your PC's own power — no queues, no server wait.
Automatic updates — the app keeps itself current.

Quer experimentar a versão mais recente?

Comece grátis na web Baixar aplicativo de desktop