TTSNative 2.1.0
TTSNative 2.1.0 is the first version you can install on Windows 11 like any other app, with the default voice included and updates delivered inside it. It also brings the guided Studio Voice session and custom voices with styles and sharing, audiobook exports with chapters, characters with their own voices, a library of 25 public-domain books with ready-made reading experiences, phones on your home Wi-Fi, background music under the narration, a pronunciation dictionary, bookmarks, search and a sleep timer.
New
- Install TTSNative on Windows 11 like any other app: it opens in its own window from the Start menu or an optional desktop shortcut, with nothing else to set up.
- The default voice comes with the app, so you can start listening right after installing.
- TTSNative lets you know when a new version is available, shows what's new, and installs it for you once you agree. You can turn these checks off.
- See what changed in every version under Settings, About, even without an internet connection.
- Play quiet background music under the narration: turn it on from the music button in the player, choose from eleven free tracks (ambient, fantasy, cinematic, lo-fi, focus and nature), and preview any of them before you pick.
- Music has its own volume, separate from the narration, and follows it: it pauses when you pause, carries on when you play, and fades gently in and out.
- Set the narration and the music volumes side by side in the new audio mixer in the player.
- Music is off at the start of every new narration. TTSNative remembers your track and volume on each device, but you always choose when the music plays.
- Organise your library: put documents in folders (nested as deep as you like), gather them in collections (a document can be in several), mark favourites, give each one a type and genres, and add your own tags.
- The library has a rail of views: all documents, continue reading, favourites, recently added, your collections and your folders, with a matching menu on phones.
- Search finds authors, file names and tags as well as titles; filter by genre, type and progress; sort by author too.
- Select several documents at once to move, tag, favourite, add to a collection or delete them together, or drag a document onto a folder.
- Poems and other verse keep their lines: a line break is read with a short pause and shown as a line break, not as a new paragraph.
- Explore: 25 public-domain books come with TTSNative (Alice, Frankenstein, Pride and Prejudice, Sherlock Holmes, Poe, Dickinson, Whitman, Douglass, Thoreau and more), grouped by kind, each with a cover, a description and content notes.
- Every included book says where it is public domain: always the United States, and elsewhere by its author's death year, naming the countries where a longer term still runs.
- Open an included book and it joins your library like any other document; hide the ones you don't want to see. The credits name every edition, cover artist and source.
- Reading experiences: six of the included books come with a curated cast. In the reader, choose **Multi-voice** to give Alice, Scrooge, Holmes, Elizabeth Bennet and the rest their own voices and to hand untagged lines such as the Raven's "Nevermore" to their speaker.
- **Standard** takes an experience out again and keeps every voice you chose yourself. Sound effects and music are described but arrive with a later version; nothing starts playing on its own.
- Record a guided studio session for your most faithful custom voice: a microphone check, short passages to read or a long recording to import, then pick the best of five candidate voices by ear.
- Choose how a custom voice delivers its lines: steady, natural or expressive.
- Clean up a voice recording with background noise reduction, a rumble filter and tone controls, with a before-and-after preview.
- Teach TTSNative how to say names and words with a pronunciation dictionary.
- Bookmark the sentence being read, add notes to bookmarks, and export all your notes.
- Search inside a document and jump to a match; search and sort your library.
- Set a sleep timer that stops between sentences, after a set time or at the end of the section.
- Adjust the reading text's size, line spacing and typeface.
- Rename a chapter or move where it starts.
- Export subtitles timed to each sentence, with the title, album and track filled in on exported audio.
- Read HTML pages, FB2 and MOBI books, and import a public web page by its address.
- Running headers, page numbers, footnotes and citation markers are skipped when reading aloud and exporting; they stay visible, marked as skipped, and you can read them all again or skip any sentence by hand.
- Fix how a word is said from the reader: select it, write it the way it sounds (gnocchi → NYOH-kee), hear both versions, and keep the fix for this document or for all documents.
- Export M4B audiobooks with chapters, a description and a cover, and read-along EPUB books whose text is highlighted as the audio plays.
- Pause and resume audiobook exports, even after a crash or restart; the new Exports page lists them, shows each one's place in line, and lets you resume, cancel, download or delete them.
- Send a finished export to your own Audiobookshelf server.
- Give characters their own voices, read quoted speech in a dialogue voice, and read chosen sentences as an announcement, a phone call, a radio broadcast, from far away, as a robot, quietly or as a memory.
- Mix two to four built-in voices into a new voice.
- Paste text with speech marks for pauses, voices, effects, styles, sounds and pronunciations.
- Share a custom voice as a file and import one shared with you; the file is checked completely and previewed before anything is imported, and an import never replaces one of your voices.
- Pair a phone or tablet on your home Wi-Fi to read and listen from it, with the computer doing the work (optional, off by default and after every restart; home networks only).
- Let other programs on this computer use TTSNative's voices, including OpenAI-style voice names (optional, off until you turn it on; Settings shows the exact address to use).
- Fine-tune a studio voice on its takes with an NVIDIA graphics card, and keep it only after hearing it against the voice as it is (experimental; in a blind test it didn't sound clearly better).
- Rebuilding a studio voice's speaking styles keeps the ones from before, and you can switch back.
- A high-contrast option, wider letter and word spacing for the reading text, and support for Windows high-contrast themes.
- The app's screens open even while the server can't be reached and say so, and a Reconnecting bar shows until it answers again.
Improved
- Local TTS is now called TTSNative, with a new look. Your documents, voices and settings carry over.
- Voice recordings tell you when a microphone sounds too bright or thin, and a studio session can correct its tone automatically.
- PDF chapters now begin at their heading, so the end of one chapter no longer appears at the start of the next.
- A shorter studio session: about 15 passages instead of 48, tiers that follow what the voice uses (Foundation, Developing, Complete), a suggested next passage, and progress that weighs how clean and long your takes are.
- Studio takes name the background noise when it matters, point out a take cut off mid-word, keep the previous take to compare and restore, and say what to do about microphone problems.
- Listening times count only what is read aloud, and follow skipped sentences and the cleanup.
- The Steady delivery of custom voices is steadier, and their audio is generated again once.
- Background noise in a studio take is advice, not an error.
- Selecting text in the reader no longer starts playback from the sentence you clicked.
- English voices read money, clock times and phone numbers the way a narrator does where the meaning is clear: £1,284.50 is "1,284 pounds 50", "at 4:45 p.m." is "at 4 45 PM", and a phone number is read digit by digit. Code, formulas, ratios and lists of numbers are read as before.
- English voices read years and abbreviations naturally too: "in 1998" is "in nineteen ninety-eight", Chapter IV is "Chapter 4", and Lt. Col., Rev., Mt., Rd., Ave., Ms. and approx. become words. The text you see never changes, and a pronunciation you saved still wins.
- Every voice card says where its voice comes from and under what licence, and that it is synthetic speech rather than a recording of a person.
- Updating keeps your documents, voices and settings, and uninstalling asks twice before deleting them.
- Release history shows one version at a time: pick a version from the list or step to the older or newer one, see when it came out and what changed, and find the version you're running marked.
- The version you're running is shown at the bottom of the sidebar and leads to About, where What's new and Release history are.
- On the website, every version has its own release notes page with links to the older and newer versions, and the home page names the latest version.
Fixed
- Pausing or cancelling an export never ends it as failed, and cancelling removes everything it made; after a crash, half-built files are tidied and finished sentences are kept for Resume.
- Custom voices keep a faint, steady room tone through pauses instead of dropping to dead silence, never louder than the background just before the pause.
- Numbers in brackets, such as array[3] or [0, 1], are read as written; superscripts in equations are no longer taken for footnote markers.
- Reading progress reaches 100 % at the last sentence that is read.
- Library cards no longer push the page sideways on phones.
- Importing a long studio recording, or a shared voice file larger than 25 MB, no longer stops with a general error.
- A file or request over the size limit is answered with a clear "too large" message instead of a general error.
- An audiobook whose title is a name Windows reserves for devices (such as NUL or CON) is exported under a usable file name.
- A paired phone's name is shown with visible characters only.
Security & reliability
- Voice model files are checked before every use; a damaged file is named, and Repair downloads only that file again.
- The speech engine's built-in usage reporting is switched off, so TTSNative sends no usage data.
- Security updates for the optional voice-cloning engine.
- Requests passed on by a proxy, tunnel or port forward are refused, and only a page opened as localhost or 127.0.0.1 counts as the person at this computer.
- A paired phone works only from the address it paired from and for 30 days at most (14 days unused); paired devices reach only an explicit list of features, and phone access turns itself off when the computer's network changes or after it slept.
- On Windows, phones can connect only on a network marked private; on macOS and Linux you confirm each time that it's a home network you trust.
- Other programs' use of the voices is off by default, needs a token shown once and stored only as a hash, refuses web pages, and has limits on length, size, rate and time per request.
- Released versions of the Windows app no longer accept the debugging switch its automated tests use, and a release can't be built with it.
- A request that reaches the computer's own address from another device is refused outright unless network mode is on; phones and tablets come in through LAN mode.