ToBe SAID - TTS Engine for iOS

Hey this sounds cool!
I know cloning is all that's currently available, but will voice design come later like Mac transfers will?
I don't have a Mac, so will maybe only get voices from other Mac users unless I use my own.

Love the default voices but wish the cloning quality voices were slightly higher. They always sound as if their muffled even when I do use very high-end recordings. Don’t know what that’s about but it might be something worth looking into as would battery drain issues. Works great just drains my battery rather aggressively and sticking it on the charger only helps so much.

I can't help but notice that when navigating the screen the TTS synth hardly pauses at all between the name of icons and voiceover hints. Is this something that will be adjusted/fixed in the future?
I like the sound of these voices though!

They really need to come out withh phones where we can read our own screen using braille. And plus, if I had my wish, I make an app called Nuf Said, where it would be the braille N speak voices, the braille lite 2000 voices plus intellitalk voices from intellitalk 1, plus the millennium 20 and yes dectalk, and I would use nuf said because some people just want to be simple

hi,

Here are my observations on iOS version. I love the default voices. Navigation when swiping left and Wright is responsive. Hopefully they will add the ability to import the voice that you designed from Mac to the iOS app soon. Other than that, not bad for a first release.

Great joke. Although I would never want to hear the voices from the Braille and speak or barely ever again. No thank you. Unless it was from the Braley M2, perhaps those wishes are better. But not the rail light 2000. I had that it was awful. Anyway, it would be pretty interesting if you could put actual Braille on a phone screen. I kinda like that idea. But I don’t think that’s gonna be happening anytime soon. As far as the app itself, I just tried it. I was really only able to type into the app itself to make the voice speak. I couldn’t figure out anything else beyond that. I couldn’t figure out how to get the voices to show up in the voiceover settings or anything. I don’t know maybe I just don’t know what I’m doing, but I really couldn’t figure it out.

Hello, just a quick question: are all Pocket TTS languages supported on the iOS version?

I know that on Android, when you select a voice, there’s an option to choose a language and download the Pocket TTS model for that language. On iOS, I can see language tags in the voice names, but there doesn’t seem to be a language selector, especially when cloning a voice.

So I’m wondering: does iOS automatically use the appropriate language based on the voice or language tag, or is Pocket TTS currently only available in English on iOS?

I hope more voices get added. Currently I can only use default and narrator. The other voices sound robotic and foreign.

Hi, I’ve got this app working. I re-downloaded it. I have the free version. I’m using the storyteller voice right now. One thing I’ve definitely noticed, but this Voice is pretty responsive, but there’s definitely a lag from swiping on the screen to when the Voice actually speaks. I’m using this on an iPhone SE 2022 with iOS 26.6.1. I mean these voices might take a little bit more processing power to work, but they seem kind of nice. So there’s four voices. I actually didn’t have to recite the phone in order for the Voice to shop voiceover. It seems to be working pretty well so far. Thanks for making it available. So does the pro version just give you more voices? I like the ones that are currently available.

I just switched the voice to the default. That voice is a lot more responsive. It’s interesting. The difference in response on swiping on the phone from one voice to the other. These voices are pretty good though. I’m going to try this on my regular iPhone, my iPhone 15 and see if it’s any different. It’s got a newer processor in it. I know these AI generated voices so maybe it takes a little bit more processing power.

Villain has a strange accent, but Storyteller is a British voice. As for voice cloning, the cloned voices are probably 8KHz or something.

The time is being spoken as whole numbers rather than time itself. For example, the time says that it is 1:20 5 PM. It speaks it as a whole number as it is 120 9 PM. This is the case on both of my devices using the default voice. It does seem to be a lot more responsive on my iPhone 15. The storyteller Voice lag a lot on both devices. I haven’t tried with narrator or villain. I’m enjoying the default voice that it’s pretty nice. There’s some interesting pronunciation issues. Those are difficult to write down how it’s pronouncing things so I’m sure that will get corrected. This is really good for the first release though.

I’ve had to go back to the vocalizer voices because there’s a pretty considerable lag from swiping on the screen to the time of the speech tells you what you’re highlighted on. I found this to be the case with the default voice and storyteller voice is even more laggy. I tried this on both an iPhone 15 and my iPhone SE 2022 and I had the same experience with both. I think those are good voices. I just don’t think that my devices have the processing power to do with them. I might have to wait till I’m on a newer device to try this app out again. But I did try with the free version. Thanks for making this up for us. I hope that people with the newer devices with a better processor can enjoy it more out of it. Those are pretty good Voice.

Thanks for getting back to us another info. That’s helpful. Yes I was using the default voice. That one did seem to be the best for navigating the phone. Are those all the voices that will be available in the free version of gonna be more Voice coming? At least you’re aware of the other bugs of things I mean, they obviously can fix whatever things you can fix at a time and that’s completely understandable. Maybe I’ll try the app again once I update iOS 27 or something but still a great app.

I paid for the pro version or at least I did for the Mac. When I tried to do it for the phone, it told me the update was free because I earned a previous version.

Just something for you guys to be aware of. I re-downloaded this app today in order to try it out again on my iPhone SE 2022. When I selected the Voice named default it completely crashed. VoiceOver and I lost speech completely. I had to get my aunt to help me get back on voiceover. Just something to be aware of it you may have the app crash and loose speech completely. The only thing that was left in voiceover was the navigation sound and I couldn’t do anything with any of the settings. I’m thinking maybe that this phone can’t handle this app because this did not happen at all yesterday, but for some reason when I went to select the voice again after I re-download completely crashed.

You’re welcome. I figured that was pretty significant of a bug. I definitely didn’t have the issue when I first downloaded the app. It was something to do with re-downloading the app and then going into the VoiceOver settings and selecting default. As soon as I double tap it select it was when it crashed.

I tried adding an audio to test if the voice cloning works properly, but when I added nothing came

As my subject says, I’m a bit hesitant to re-download this app unfortunately. But that’s like based on my complete crash of voiceover speech when I did that on Saturday. I didn’t happen when I first downloaded the app. It was only after a re-download and then selecting the default voice. But yeah it’s definitely made me a bit hesitant to re-download this again. I ended up just like deleting the entire app on both of my devices. I know you said you’re going to work on fixing that for the next update. I like the voices. I just don’t think that at least in my phone for the SE 2022. I don’t think maybe that processor can handle those kind of voices as well because that was a phone that it crashed on. So yeah. It does want a bit better on the iPhone 15. There’s still a little bit of lag navigating the phone with this app, but it’s definitely better with a default voice with storyteller. I know why that is now though. Thanks for working on this app.

It’s OK, I know you’re fixing it. Unfortunately I don’t have any e-book reader speech center or anything but thanks for the suggestion.

Thank you so much for the recent improvements to the app. It works great with AirPods now, and I can finally clone voices without any issues. It completely fixed the problem from the last update. The only remaining thing I've noticed is the speech volume, but that’s a separate matter. I'm just glad the app is stable now.

Yes, the latest update is great! The cloned voices are not 8KHz-ish now. I'm looking forward to being able to use the app in multiple languages, and clone new voices in all the languages I speak for use with VoiceOver.

1. Does the pro version let us clone multiple voices? Does it let us rename them?
2. Can you add the ability to sync voices between multiple devices, and share them with other users?
3. Why do we have to download the pro version as a separate app rather than an in-app purchase?

Sorry, but I find this particularly important, to the point that it may urge many users to upgrade to the premium version, and while relevant, this post touches upon a separate issue, so I wanted to create a separate post instead of editing the previous one.
According to the list ChatGPT gave me, the official Pocket TTS release from Kyutai already supports six languages:

• English
• French
• German
• Italian
• Portuguese
• Spanish

ChatGPT also stated that multiple community-trained checkpoints beyond the official model exist, with support for more languages including Czech (vvolhejn/pocket-tts-czech), Hindi (saryps-labs/pocket-tts-hindi), Korean (seastar105/pocket-tts-korean-300m), Nepali (himalaya-ai/pocket-tts-nepali-6l), Persian (mehdi-hf/pocket-tts-farsi), Russian (amekhrishvili/pocket-tts-ru), Telugu (prasadvittaldev/pocket-tts-telugu-female-syspin), and Turkish (kaanhgunay/pocket-tts-tr). Note that ChatGPT initially listed four languages only (Czech, Hindi, Korean, and Persian), and provided the others only after I explicitly prompted it to check for these languages as well. Disclaimer: I didn't verify any of the information ChatGPT provided, nor do I know whether these community-trained checkpoints match the model you bundle within your app, even if they exist. Here's the source URL that ChatGPT referenced.

Neither the App Store description nor your post where you introduce the Mac app mentions any of the supported languages. As for the community checkpoints I shared with you earlier, I request that you check all of them one by one and include any suitable ones in the app. Restrictive licenses may prevent you from using a non-commercial asset in a commercial app, but quality and naturalness is a separate thing. So I'll be honest with you: Please do not dismiss these community models all at once; some of them may be decent and legally possible to ship with the app.

It's great that the app supports several languages beyond the official ones, but Turkish is not among them. If it's because you don't consider kaanhgunay/pocket-tts-tr good enough, let me test it out as a native speaker of Turkish. If it's because of legal concerns that you avoid including it, it's licensed under CC By 4.0, meaning you can definitely use it commercially. If there's anything else, I'll probably have to acknowledge it. I think I addressed these two points I mentioned though.

I’m not sure if I had the new version of the app or not but last night it said re-download the app so I did and then went to select the default voice and then I got a complete speech crash again. One of my side friends go in your voice for me for voiceover and I’m back on the regular built-in phone voices. I’m not sure how to tell which version of the app I’m running honestly and I really don’t even know how to update apps but I would try again if I know if I have the right app. It’s a free version and an iPhone SE 2022 running iOS 26.7 thanks I didn’t even say like selected default or anything it just as soon as I double tapped on it it completely turned off the speech.

How would I be able to install this on Mac? The App Store is completely and utterly broken. I can't install or update any apps, I mean, actually that's a bit of a lie. I can update apps I was being a bit overdramatic there but yeah I can't install or re-download any apps on the App Store

It’s OK. It could be me too I don’t really know. Like I said I’m not exactly sure how to tell what version of the app I have and I’m also really not sure how to update apps so I mean the old one could’ve been re-downloaded. Because my search the name in the App Store it says re-download didn’t say anything about a different number or anything like that so I don’t know. I didn’t do this on my iPhone 15. I think it’s gonna be something with the processor in the SE. He did say tested it with an older 10 something so I don’t know. It’s really interesting when you’re able to happen then you literally get no voice at all you’re fixing I know that.

It could be a device-specific issue or a system bug, because I've had similar problems with other TTS engines on the same device, so ToBe Said may not be the culprit here.
Having said that, I have a couple more questions:
Thanks Bakhtiar for the explanation; now I get why you can't add that Turkish model. I really want to support development and help you fund required hardware purchases to train new languages. I've just bought the pro version; the download is still in progress as I'm writing this post. I request you to introduce support for Turkkish and other languages though, including Central Asian Turkic dialects and Arabic. Also, keep in mind that you don't necessarily have to own the hardware yourself; you can look for other online methods like Google Colab or Hugging Face's own training infrastructure. Finally, I think Pocket TTS has more default voices other than the four (Default, Narrator, Storyteller, and Villain). Does the 100M variant have all of those or are they only available in the 300M model? Can you add them to ToBe Said?
PS: You can actually add the abillity to design voices to iPhones with at least 8GB VRAM, or if that's not possible, you can just use your own mac as a server, let us upload a description, generate the voices on your own device, and then let us download them for a small fee.

Just launched the pro version, but what's Share Mode supposed to mean?

Interesting that it works on your iPhone SE 2022. That’s exactly the same device that I’m having issues with for this particular TTS. I have no idea why. I’m thinking that maybe I didn’t get the updated version of the app because when I went to the App Store said to re-download and I did delete the app before it crashed like that. I don’t know. I honestly don’t know how to find updated apps or how to even update apps so it could very well be something like that. But it was working fine on my iPhone 15 so yeah I don’t know but it’s all gonna get sorted out eventually.

Here are a few:

1. Seven-year-old girl (a cheerful/cute-sounding seven-year-old girl still learning how to read and write, trying to read slowly but with childish enthusiasm and excitement)
2. Seven-year-old boy (a cheerful seven-year-old boy still learning how to read and write, trying to read slowly but with childish enthusiasm and excitement)
3. Old man (a deep, calm, warm, and friendly old man in his seventies)
4. Voice actor (a deep but clear, low-pitched, confident and expressive man in his forties)
5. Casual (a natural, warm, sincere, and modest man/woman in his thirties speaking fluently, without too much exaggeration)
6. Commander (a military commander, as if addressing his soldiers in a solemn, passionate and determined tone)
7. Male news reporter (a high-pitched, confident news presenter with clear articulation and somewhat exaggerated intonation)
8. Female news reporter (a dramatic, high-pitched, confident news presenter with clear articulation and somewhat exaggerated intonation)

Or how about we be able to import voices generated using Qwen3-TTS?
My name is Enes by the way.

There’s already an explanation in here as to why there were four voices for the free version. I literally cannot remember what it was now, but there was a really good explanation from the developer. Look I’m happy with the four voices. As long as I can get it working and everything it’ll be fine. I guess if we really want more voices, you can see if maybe the pro version offers more. But I’m just gonna say it feels a little unfair. Got some people seem to expect so much from this one app instead of understanding why there’s things that are certain way for it. Suggestions are cool yes but sometimes the suggestions come across like people are expecting things.