By VestraCode, the developer of ReadMe.
Offline TTS Models for Android EPUB Reading | ReadMe
ReadMe by VestraCode is an EPUB and PDF reader with offline text-to-speech for Android. Download Supertonic 3 directly in the app or import a compatible Piper, Kitten, Kokoro, Matcha or VITS model to listen with local neural voices.
OFFLINE NEURAL TEXT-TO-SPEECH FOR EBOOKS
Choose a local voice that fits your book and device
ReadMe combines EPUB and PDF reading with configurable text-to-speech on Android. Its ReadMe Local TTS workflow supports compatible sherpa-onnx neural model packages, so downloaded voices can synthesize supported book text on the device without a permanent internet connection.
Supertonic 3 in the app
Download the multilingual model directly from ReadMe, then choose from 10 speaker IDs.
Piper voice choice
Import a compatible Piper voice for fast local narration in its supported language and accent.
More neural families
Compatible Matcha, Kitten, Kokoro and VITS/Inflect packages expand the available local choices.
Supertonic 3 at a glance
Languages31 supported languages
Speakers10 speaker IDs (0-9)
Sample rate24 kHz
ProcessingLocal and offline after download
InstallationDownloadable directly inside ReadMe
Reading formatsTTS for supported EPUB and text-based PDF content
Supertonic 3 is the easiest place to start
Supertonic 3 is ReadMe's featured local neural TTS model. Instead of searching for separate voice archives, you can download its optional module directly in the app, select a supported language and try one of its 10 speaker IDs. Once installed, speech generation runs locally for offline listening.
The official sherpa-onnx model page lists 31 languages, 10 speakers and a 24 kHz sample rate. The same multilingual package covers the supported languages below. Open the Supertonic 3 model page and voice samples.
ReadMe also lets you adjust the narration speed and continue supported playback with the screen off, making it useful as an offline EPUB reader with text-to-speech for walks, commutes and hands-free listening at home.
Languages supported by Supertonic 3
EnglishKoreanJapaneseArabicBulgarianCzechDanishGermanGreekSpanishEstonianFinnishFrenchHindiCroatianHungarianIndonesianItalianLithuanianLatvianDutchPolishPortugueseRomanianRussianSlovakSlovenianSwedishTurkishUkrainianVietnamese
The language names above correspond to the official codes en, ko, ja, ar, bg, cs, da, de, el, es, et, fi, fr, hi, hr, hu, id, it, lt, lv, nl, pl, pt, ro, ru, sk, sl, sv, tr, uk and vi. Voice character and pronunciation can vary by speaker and language, so test a familiar passage before a long listening session.
Which offline TTS model family should you try?
The recommended first choice in ReadMe: one directly downloadable multilingual package, 31 languages and 10 speaker IDs. It is designed for local neural synthesis after installation.
A broad collection of efficient single-language voices. Download the compatible voice you want, import it into ReadMe Local TTS and use it offline. The English catalog includes British and American voices in low, medium and high variants where published.
Compact English neural models are available in nano, micro and mini variants. The catalog includes FP16, FP32 and INT8 packages, giving users options with different package and device requirements.
Kokoro is another compatible English neural TTS family available through the sherpa-onnx model catalog for local use after import.
The English catalog also publishes Matcha/Icefall LJSpeech and compact VITS/Inflect models. Import a compatible package and test it on the target Android device.
How to use an offline model in ReadMe
01
Choose the installation path
Download Supertonic 3 inside ReadMe, or obtain another compatible package from its official sherpa-onnx model page.
02
Install or import the model
Install the in-app module or import the compatible local archive from your Android device.
03
Select language and voice
Choose the installed model, its supported language and an available speaker, then test a short passage.
04
Listen offline
Open a supported EPUB or text-based PDF, adjust speed and continue local playback without a permanent network connection.
Model compatibility and performance
Every model linked below is an offline sherpa-onnx TTS model supported by the ReadMe Local workflow described by VestraCode. Download or import is required before offline use. Model size, synthesis speed, memory use, available speakers, language and pronunciation depend on the selected package and Android device.
Supertonic, Matcha, Kitten, Kokoro and VITS model pages
Piper British English voice pages
vits-piper-en_GB-alan-lowvits-piper-en_GB-alan-mediumvits-piper-en_GB-alba-mediumvits-piper-en_GB-aru-mediumvits-piper-en_GB-cori-highvits-piper-en_GB-cori-mediumvits-piper-en_GB-dii-highvits-piper-en_GB-jenny_dioco-mediumvits-piper-en_GB-miro-highvits-piper-en_GB-northern_english_male-mediumvits-piper-en_GB-semaine-mediumvits-piper-en_GB-southern_english_female-lowvits-piper-en_GB-southern_english_female-mediumvits-piper-en_GB-southern_english_male-mediumvits-piper-en_GB-vctk-medium
Piper American English voice pages
vits-piper-en_US-amy-lowvits-piper-en_US-amy-mediumvits-piper-en_US-arctic-mediumvits-piper-en_US-bryce-mediumvits-piper-en_US-danny-lowvits-piper-en_US-glados-highvits-piper-en_US-hfc_female-mediumvits-piper-en_US-hfc_male-mediumvits-piper-en_US-joe-mediumvits-piper-en_US-john-mediumvits-piper-en_US-kathleen-lowvits-piper-en_US-kristin-mediumvits-piper-en_US-kusal-mediumvits-piper-en_US-l2arctic-mediumvits-piper-en_US-lessac-highvits-piper-en_US-lessac-lowvits-piper-en_US-lessac-mediumvits-piper-en_US-libritts-highvits-piper-en_US-libritts_r-mediumvits-piper-en_US-ljspeech-highvits-piper-en_US-ljspeech-mediumvits-piper-en_US-miro-highvits-piper-en_US-norman-mediumvits-piper-en_US-reza_ibrahim-mediumvits-piper-en_US-ryan-highvits-piper-en_US-ryan-lowvits-piper-en_US-ryan-mediumvits-piper-en_US-sam-medium
An offline AI voice reader is still an ebook reader
A useful Android text-to-speech reader should do more than generate audio. ReadMe keeps the listening workflow beside chapter navigation, bookmarks, highlights, saved quotes, full-text search, reading progress, collections and favorites.
You can read an EPUB on screen, switch to a local neural voice, change the speaking speed and keep listening in the background. ReadMe also supports PDF, CBZ and CBR, while spoken PDF playback requires extractable text.
This combination makes ReadMe relevant when you need an EPUB reader with TTS, a PDF read-aloud app, a multilingual ebook reader or a way to turn supported files into locally generated spoken audio.
Can ReadMe read EPUB books aloud offline?
Yes. Install Supertonic 3 or another compatible local TTS package first, then ReadMe can synthesize supported EPUB text on the Android device without a permanent internet connection.
Which offline TTS model is easiest to install in ReadMe?
Supertonic 3 is downloadable directly inside ReadMe. It includes 31 supported languages and 10 speaker IDs in one multilingual package.
Does ReadMe support Piper voices?
Yes. Download a compatible Piper package for the required language or accent, import it into ReadMe Local TTS and test it with a familiar passage.
Can I use Kokoro, Kitten, Matcha or VITS models?
ReadMe supports compatible sherpa-onnx packages from these local neural TTS families. Use the linked official model page to obtain the package and import it into ReadMe.
Do all listed voices work without internet?
The listed local models synthesize offline after their files have been downloaded or imported. Downloading the package initially requires access to its source.
Can ReadMe read PDF files aloud?
ReadMe supports TTS for text-based PDFs when extractable text is available. Image-only scanned PDFs may require OCR before text can be spoken.
Where can I browse models for other languages?
Use the complete sherpa-onnx TTS catalog linked below. Each model page identifies its language, samples and download package.
Choose your ReadMe offline TTS setup
Start with the directly downloadable Supertonic 3 module for 31 languages and 10 speakers.
Use Piper when you want a specific efficient language, accent or published quality variant.
Explore compatible Kitten, Kokoro, Matcha and VITS/Inflect packages for additional English neural voices.
Test model speed and memory use on the Android device before committing to a long book.
Keep reading, library organization and offline spoken playback together in ReadMe.
Bring a local neural voice to your next book
