Announcing the first public beta of Runa TTS for iOS: bringing the high quality, natural sounding voices from ReadSpeaker to VoiceOver and other apps that rely on Apple's speech synthesis API's.

I was very disappointed to learn these voices aren't the classic ones we have on Windows. I understand this isn't Adi's fault, but rather the company it's licensed from which is a shame, and there's probably no way to fix it. I hope people find it useful, but I won't be purchasing this unless we get the same high quality voices. To be clear, I have no problem paying, but I don't want inferior voices that only take up about 20 MB at the highest quality when the legacy voices were a few hundred megabytes and sounded significantly better as a result. I'm not sure why companies are going down this path of significantly reducing file sizes and sacrificing quality in the process, particularly when modern devices have ample amounts of storage space and processing power, but it is what it is. I wish the best of luck to Adi and the team. The only good "neural" voices I've heard are from services like Eleven Labs, but those can't run offline.

Hi there, so I set up a voice profile, call it Welsh, then added the Welsh voice. It’s not in my VoiceOver settings. I do have a show in voices rotor setting turned on for that profile.

Could you add a preview button for the selected voice before downloading?
I want to use your automatic switching voices.

Enjoying beta three so far. My only personal complaint is that the name of this website is pronounced as my subject line, when using UK English or Australian English voices. When I use US English voices, it pronounces properly. Really enjoying Jack (enhanced) Australian voice at the moment, so it's a small price to pay for a decent Australian voice, I guess. 🤷

The app is working well so far. I have voices Aaron and Kate enhanced on the rotor. Thank you for the update regarding reading long streams of text. I realize this app is still development regarding that. There are some instances where the screen would dim after reading long streams longer than 30 seconds. again the app is still in development. Also regarding pronunciation, with an address with a capital letter in it i.e. 301N X St. it pronounces the capital N as Newton rather than north and street as a saint. Only if the Aaron and Kate enhance voices behave this way. None of the Apple voices behave this way. I realize this is a very minor. And could probably be fixed and pronunciation settings.

I could be wrong here, but it seems like Runa TTS only follows the pronunciation dictionary of VoiceOver for the default language. For example, I am in the US, so my default is US English. Anything I have in the pronunciation dictionary works fine with the US English voices for Runa TTS., For example Apple VIS, or AppleVis (note the spelling of both of those), what read as, "AppleVis".
However, when I switched to UK English, or Australian English voices of Runa TTS, the voices are great, but seem to completely ignore the pronunciation dictionary.

Appreciate everything you're doing. 😊

Hello.
I went to manually add it and it wasn't there. the trick was restarting my phone.
Now I have 23 pages of languages, over 200 of them, most of them things I have never heard of, all of them having both access mind voices that I have installed in them.
Like voices that don't even match their languages, for example faroese, ainu, etc. all of which of course have no voices but have the 2 accessmind voices in them. i only expected 1 new language entry, welsh, like in beta 1. i got 'em all. haha.

Hi all,
Kyler here. Overall, Runa TTS is still working great. However, I have noticed in beta 3 that the pauses between sentences are shorter, and the setting in Advanced settings to increase it by 200ms doesn't seem to work even if I set the default profile voice to the one I'm using (Julie). Can anyone else confirm whether they're seeing this?

Hello,
Is there a plan to add a feature to reduce pauses between sentences?

sorry for all my issues. i've figured it out. appreciate you doing this! i'll definitely be buying it when it comes out.

Just curious: How are you getting voices for Faroese and Ainu? None of those are showing up for me. I know you said you had AccessMind voices for all of them.

Hey, the third beta works well with me, as I am catching on to the comments.

Thanks for everything!

For people that are saying that you don’t have the same voice quality. I just wanted to share a little quick FYI for you. All of the current voiceover voices that are available are not the same quality as the ones in the computers in the Apple gives us. The ones that are called premium are definitely not the same qualities as premium high in windows. I can’t speak to the Mac because I don’t own one. So that’s really not necessarily gonna be able to be helped with this particular app either. Just more of FYI to everybody. :-) I think this has to do with the processing power or phones versus computers. They’re capable of larger file sizes but I think all of the companies that are giving licenses or invoices are making smaller sizes. It’s not just this one. :-) Hope you guys enjoy this app though. I’m going to stick with the available voices itself. Although I did re-download the apps. Even the quality of those is not the same as what’s in the computer. I know because one of my friends showed me with her computer running windows at NVDA so it’s definitely different even with that. I hope you guys have fun with this app though.

The beta 3 is working perfectly for me;
Setting voice profiles in my voice rotor is great, especially in iOS 27 beta, where the automatic language switching is messed up a bit.
Also, I can't wait to hear about the NVDA addon version of this app.
I hope you would make us a forum for that, also.

Hello @Adi Kushnir, I really appreciate what you and the team are doing to bring additional voices to iOS.
One pricing suggestion: please consider supporting iOS Family Sharing. Either make the initial in-app purchase eligible for Family Sharing or offer a separate family version.
For example, if the individual version were $30, I’d gladly pay $50 for a family version that could be shared with my family. I think this would be a great option for households with multiple Apple users.

There should be no up charge for a family version. Those are the kind of money grabs that will cause me to not purchase the product.

Hello Adi kushnir,

This app works really well so far. The only problem that I’m currently having is there is some popping here and there when a voice starts to read something long. For the voices, I’ve tried, it only happens with Beth and Julie, which I both use very frequently. also, will you guys be adding more English United States voices in the future such as James? Something that has came up to me in terms of When The public beta testing ends is that I know we can use the beta for at least until it expires, but could there be a discounted or free like migration or anything like that to the final App Store release so then you wouldn’t lose any voices or profiles you have set up. I got this app and I’m loving it, but I don’t want to have to re-get it in the future for any further cost than what I got originally from if you know what I mean. But because the voices belong to a paid company I don’t think that will probably happen. Or at least have it so whatever voices we have now we can keep but if you want to purchase any new voice, you would have to pay for it. But then that’s money coming out of your pocket, which I don’t want either.

Thank you for understanding!

Although beta 3 appears to have improved the issue of VoiceOver stopping when the screen locks while reading longer texts, I would still like to point out that the problem has not been completely resolved, as the reading can still be interrupted in this situation.

I would also like to report another issue affecting the French voices. When using one of the French voices to read text letter by letter, accented characters are not pronounced correctly. In particular, when the voice encounters an “É” or an “È,” it simply pronounces the letter “E” without the accent. This issue is still present and affects the way accented characters are read aloud.

Hello.
Everyone here has been talking about DECTALK as the new TTS engine, but I don’t agree.
If it were possible to add Microsoft’s natural-sounding voices to iOS, Android, and Windows platforms, I’d buy them without a second thought.
Microsoft supports hundreds of languages and over 400 voices, and all of them can work completely offline right on the device.
I’d really love to be able to purchase and use these under a license.

I’ve been following AccessMind’s work since 2024.
I previously purchased the Cerence TTS voices they released under the name “Tiflotecnia Voices,” and I’m still using them.
The product is priced at 90 euros and includes a package with 65 languages and nearly 170 voices.
Offering such a comprehensive package at a reasonable price of 90 euros has more than proven to me just how reliable and honest this company is.
As a visually impaired person living in Turkey, even though this price is a bit expensive in my country, it’s still quite reasonable.
Now I’m asking Dennis: exactly where is the “money trap” here?
Some people really complain so much that they expect every product to be free.
Instead of constantly thinking only of yourself and whining when you leave a comment, try to consider the people who are selling the product and trying to make a living from it.
Stop complaining about everything for no reason—just enjoy the product you’ve been offered.
Accessmind isn’t my dad’s company, but I support them wholeheartedly because of the great work they do.

The Microsoft voices are certainly doable.
I've seen it done in apps like Sonarpad, and it's free.
But you can only use it for reading, you can't use it with Voiceover.
But if you just want to read a document or book, Sonarpad is the app that I know of that can currently do it on iOS.

I amusing the Erin voice, if I try to read a string of numbers, speech just totally drops out. Other than that, I amreally loving this voice. Unfortunately I can't really justify paying 20+ dollars when I am only using one voice, but I will just enjoy using it and help to report any bugs that I can while the beta is still going on.

Hi and welcome.
Thank you for developing this.
I would like to draw your attention to an error in the pronunciation of some individual Arabic letters, for example when writing a message.
These errors do not appear when pronouncing words and sentences:

" ة " nothing is pronounced at all, and it is supposed to be pronounced as " تاء مربوطة Ta' Marbutah".
" ى " is pronounced as "ااا aaa" and the name of the letter is "أَلِفْ مقصورة Alif Maqsurah".
All short diacritics are not pronounced (" َ " فتحة Fatha, " ُ " ضمة Damma, " ِ " كَسْرة Kasra, " ْ " سُكْون Sukun, " ً " تنوين الفتح Tanwin Fath, " ٍ " تنوين الكسر Tanwin Kasr, and " ٌ " تنوين الضم Tanwin Damm).
Thanks and regards.

Dear ReadSpeaker / Runa TTS Support Team,
I am writing to report an issue regarding voice recognition in iOS VoiceOver after installing a voice through your app.
Here are the details of the problem:
* Device / OS: iOS
* App: Runa TTS with ReadSpeaker
* Downloaded Voice: Polish - Aneta
* Issue Description:
I have successfully downloaded and installed the Polish voice "Aneta" within the Runa TTS application. However, when I navigate to iOS Settings > Accessibility > VoiceOver > Speech > Add Voice to Rotor (or Voice Settings) and select Polish, the downloaded voice does not appear in the available voice list at all.
Restarting the device and re-downloading the voice package did not resolve the issue. VoiceOver seems unable to index or register the installed voice from the app.
Could you please look into this compatibility issue or provide guidance on how to properly activate the voice for VoiceOver?
Thank you for your support and assistance.
Best regards,
Kuba

well, I am so glad that you guys doing it.
and I am using it really well.
it has 2 Chinese Hong kong voices.
and the english voices also really cool.

I first want to say thank you for developing this application. I am currently using the beth voice at the moment. However, one of the bugs that I am noticing is that the fact that beth is saying "applevista" instead of applevis.

this + voxlibri is a killer combo!!!
Let's you read books with these voices.... on device. holy crap I'm amazed. that's all I wanted to say.

I'm not using iOS, but this is just my opinion. The fact that he's offering a one time purchase for these voices is good, giving the cost of apple developer account per year. It is the beta software, you don't see the final price yet. Give the guy a break. Complaining gets some of you nowhere.

I am absolutely thrilled to have some of these voices available in VoiceOver. They don't even act weird and sluggish like many add-on TTS systems. I love Bridget. VoiceDream's Bridget only sounds good when using earbuds, but has been my book narrator of choice for many years. Now I can read from Kindle and Newsline with Bridget from Runa TTS. Earbuds aren't even required. This is practically life changing for me, since I've hated the Vocalizer voices since they changed in iOS version 16.

Yes, I have noticed that attention to the pronunciation dictionary is very inconsistent. It would be nice to have that fixed. Voice samples would also be good, even though it is really hard to get a clear idea of how a voice really behaves from a short sample.

Please, consider adding Runa TTS for Mac because I'm tired of keeping a very outdated Mac OS version just to keep a decent voice, which is the old Vocalizer Alison.

I add my request for a one-time price. If we need to pay something again later because updates are needed to keep the app working with future iOS speech system changes I will understand. I just haven't noticed better support from any company that charges a subscription fee and I know how much I can afford now, but can't predict the future.

Hello, and thanks for bringing these voices to iOS.

I'm a blind user in Japan, and nobody seems to have mentioned the Japanese voices yet. The ReadSpeaker voices are old friends here: Misaki is the default voice bundled with PC-Talker, by far the most used Windows screen reader in Japan, and Sayaka is built into the Plextalk DAISY players. Having them on iOS at last means a lot to us.

I've emailed Adi the details of what I've run into, but one thing might be worth mentioning here since it may not be specific to Japanese: speech cuts out briefly near the start of a long string when the VoiceOver rate is 75% or above. At 70% the same text reads smoothly. I first noticed it reading the timeline in the X app, and it happens with both Misaki and Sayaka, though sometimes with only one of them.

If anyone else is testing at high speech rates, I'd be curious whether you hear the same thing.

Hello,
The Turkish voice Ozan, when it encounters a string in English, tries a lot to read it in an English accent but mostly fails.
Can there be an option made so that voices don't attempt to read stuf in accents outside of their own language?

1. The voices can't spell individual letters correctly. They shouldn't treat single letters as full words or abbreviations.<!--</>-->li>
2. The Turkish voices add the word "saat" before the actual time when announcing the time, which means something like "o'clock". This shouldn't happen either.
3. We should have more voices in more languages and dialects, and more personas like child or old person voices, and more emotional styles.
4. The standard voices should be clearar at high speech rates, and the enhanced vvoices should be more natural. If such a small difference in size can make a huge differences, then you should give us somewhat larger but much better voices without having to increase the size so much. You should also include the old non-neural voices.
5. You should work on a full TTS runtime that can run small models like Supertonic-3, MOSS-TTS-Nano, and Kokoro TTS, and allow real-time synthesis by streaming speech instead of generating whole audio files.