Hello everyone,
I'm an independent developer. For the past year I've been building an app called EbookVoice. It takes a book you already own, as an EPUB, a Word document or a DAISY zip, and turns it into a narrated audiobook you can listen to like any other. I'm looking for VoiceOver users who would be willing to test it properly and tell me where it falls short.
I want to be straightforward about where things stand. The app is already on the App Store, and the core of it, playing a book with background audio and full lock screen, Bluetooth and CarPlay controls, is free and will stay free. There are paid tiers for people converting a great deal of material, but listening itself is not behind them. I'm not posting here to sell anyone anything. I need people who actually depend on VoiceOver to use it and tell me what is wrong with it, because there is a hard limit to what I can find on my own.
Some specifics, so you can judge whether it is worth your time:
It converts EPUB, Word .docx, and DAISY zip files. DAISY is supported specifically because a great many BookShare titles are not EPUB.
There are 27 languages and over 200 voices. For text that switches language mid-sentence, the multilingual voices handle the switch rather than reading the second language phonetically in the wrong accent.
There is a follow-along mode that highlights text as it is spoken. I changed how that text is exposed to the accessibility tree so a braille display receives a whole paragraph as a single element rather than one word at a time, which previously made continuous reading on a display close to impossible.
There is a built-in catalogue of public domain books, so you can start listening straight away without having to find a file first.
Conversion runs on the server, so you are not waiting on your own device to finish a long book.
Much of what is in the app came from a blind tester who has been using it since the spring and told me plainly what was wrong. That is the kind of feedback I am asking for more of.
What I would most like tested in the next build:
The path from choosing a book to starting the conversion. I know from my own data that this is where people are getting lost, and I need to hear what it actually sounds like on the way through.
VoiceOver across the whole app. Anything read in the wrong order, unlabelled, or not read at all, however small.
Follow-along on a braille display, if you use one.
Long books. Very long titles are the weakest part of the system and I would rather hear it from a tester than discover it later.
If you would like to join, email me at youmbidev@gmail.com and I will add you to the TestFlight group. I am asking for an email rather than posting a public link so that I know who is in and can actually answer you directly.
I will be reading this thread and replying to everything. If something does not work for you, I would rather hear it plainly than not at all.
Thank you,
Yohan
Comments
More details please then I might join
Hi.
This sounds interesting.
I just want to ask a few questions to decide if it is of interest to me.
I just don't want to join then not actually use it that much to test because it's not something I would use daily anyway.
1. Are the voices just the system phone voices or AI ones?
2. If they are AI voices which model powers them?
3. Does it convert it into audio drama style where there's a voice for each character and narrator? or does it just read it like an audio book?
4. What if any are the limmits as in is there usage limmit for a free plan and less limmits in the paid plan?
5 Could you give me some insite as to the direction you're going with this?
I'm not really asking for an exact feature list but I just mean what is the result that you want this app to be and do well?
Thanks for that.
in response to Amrit the blind constent learner
1 & 2. Voices. Neither system voices nor a "clone this person's voice" AI model — they're Microsoft's neural TTS
voices (via edge-tts), the same engine behind things like Edge's Read Aloud. So: real neural TTS, not the flat
on-device voices, but also not an LLM narrating or an ElevenLabs-style generative voice. Over 200 voices across 27
languages, including ones that handle a language switch mid-sentence without falling into the wrong accent.
3. Audio drama vs. straight narration. Straight narration, honestly. One voice reads the whole book — you can assign a
different voice per chapter if a book needs it, but there's no per-character cast. I actually had a half-built
feature for that ("Emotional Director" — automatic expressive/per-passage voice switching) and pulled it out of the UI
because it wasn't good enough to ship. Might come back to it properly later, but right now it reads it like an
audiobook, not a dramatization.
4. Limits. Listening itself has no limit on either tier — background play, lock screen/Bluetooth/CarPlay controls,
follow-along are all free, permanently. The only metered thing is converting a new book: free accounts start with a
handful of conversion credits and gain one more every month (it accrues rather than resets), plus a per-book length
cap that covers a typical novel. Paid tiers raise both numbers, plus a one-time option instead of a subscription. If
you're mainly converting a few books you own and re-listening, the free tier alone will likely cover it — that's
deliberate, not a teaser.
5. Where this is going. Honestly, it started as a personal itch — I wanted to listen to books Audible either doesn't
carry or has no narrated version of, and I already owned them, so I built something that would just read them to me.
This is the first thing I've built at this size, and I'm now opening it up past "solves it for me" to see what it
needs to actually work for other people — which is a big part of why I'm here asking rather than assuming. There's no
fixed feature roadmap I'm chasing; I'm adding what actually turns out to matter once real people use it, and
accessibility testers finding what's broken is exactly that kind of signal.