Voice Delta — Generador de voz IA gratuito
Novedades

Voice Delta Registro de cambios

Todas las actualizaciones del estudio web y la aplicación de escritorio para Windows — en palabras sencillas, sin jerga técnica.

Aplicación de escritoriov1.8.0August 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.
Aplicación de escritoriov1.6.0August 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.
Aplicación de escritoriov1.5.0August 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.
Aplicación de escritoriov1.4.0July 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.
Aplicación de escritoriov1.3.0July 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.
Aplicación de escritoriov1.2.0July 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.
Aplicación de escritoriov1.0June 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.

¿Quieres probar la última versión?

Empieza gratis en la webDescargar aplicación de escritorio