Modes Store/Share

I would love to share the modes that I've built with others and some of my learnings and as well see the modes other people have come up with. Call it the Mode Store if you will.

🐞Failed to transcribe🐞 ? Use the new version

If the version you currently on is not working Please download the latest Download BlabbyAI Windows v1.2.0 To confirm you're using the latest version, check the top left corner of the settings window and make sure you see the version 1.2.0 there.

Phone keyboard app

It would be amazing to have a keyboard app where you can type normally and also dictate text conveniently. It would be even better if it had different models you could prompt beforehand, just to easily write custom emails on the phone or have a fun conversation with friends where emojis are used automatically.

Thank you to the Develop of Blabbyai

Just a short note of appreciation and you have done a fantastic work. Let us know when this will be available in in Iphone and Android - I am willing to sign up and pay.

Issues with not transcribe

I have had almost 25% of voice notes today not transcribed and have errors upon attempting to re-transcribe three times. Very frustrating as I just paid for a month and don't want to have to look for other solutions.

Windows desktop beta toolbar interferes with taskbar

I might be wrong about this, but I tested it in a few different ways. It seems like if I have my taskbar set to automatically hide and Blabby is open, it prevents the taskbar from showing when I have any other app open full screen or across my whole window. So for instance, with Blabby on and enabled and my taskbar is set to automatically hide, If I open up Chrome and maximize that window, I can't get to my taskbar unless I minimize Chrome.

Know which model is being used for transcription

I had this idea about allowing users to choose different transcription APIs, like on cursor when you can choose different models … it would be interesting for advanced users that know subtle differences among them (such as, some are better at interpreting parts of a sentence that are spoken in a different language mid speech … some are better at translating tone to text, some are faster, etc)

What should I do if I run out of the free 10,000 credits?

So, as far as I understand, the program is currently in active development, which is why you've made it free. But what about users who are already paying for a subscription to the browser extension of your product? For instance, I predict that this free limit of 10,000 credits will run out in about one and a half to two weeks for me. How will I continue to use your product then? Will it automatically switch to credits from the paid subscription once the free credits are used up? How is this functionality implemented? Or perhaps it would make more sense to give an option that if a user is already paying for a subscription, they wouldn’t need the free credits. What do you think?

Can't properly transcribe Claude or ChatGBT

It generates every ridiculous variation except for the correct spelling CLAUD ChetGBT Clawed Ciao, GBT Come on, it’s not that hard.

Google Login

Hello, I’m having trouble logging in with my paid account, or even when I do manage to log in, it doesn’t show my plan.

The "GPT-4o-transcribe" model is set as the default for transcription.

I'd like to see an option in the software setup to set a different default model for transcription, just like what's done in the browser extension. Specifically, I'd love the ability to choose the “GPT-4o-transcribe” model, as it's much better at processing speech than the “Whisper” model.

Allow mouse clicks in shortcurt

I really wish you would create a feature that allows me to set my shortcut to a mouse click. For example, I’d like to click the scrollwheel and have it work to activate. It shouldn’t be limited to just the keyboard. Thanks, great product, i use it everyday

Kyrgyz voice input is currently available only in ChatGPT. Other major companies, including Xiaomi, Google, Apple, Windows, and Linux, still do not provide full support for Kyrgyz voice input. In Kyrgyzstan, a country of more than 7.5 million people, there is a genuine need for this feature. At present, many of us are forced to use Russian for voice input. We strongly urge developers to add support for the Kyrgyz language. In today’s world, a language is much more than a means of communication — it is an essential part of a nation’s culture, identity, and informational independence. Recent events, including the situation in Ukraine, have demonstrated how important and sensitive the issue of language and its role in society can be. If Kyrgyz is not actively developed and fully supported by modern digital technologies, Russian will continue to dominate the digital environment, while Kyrgyz may gradually be pushed aside. Supporting Kyrgyz voice input is therefore not merely a matter of convenience — it is an important contribution to preserving, developing, and securing the future of our language.

The menu gets blocked when the browser or some other software is running

During testing, I found an issue where, on Windows 11, if you use the menu hiding feature, the menu unfolds without any problem on the desktop. However, when, for example, Google Chrome is open, your program blocks the menu from opening. You can see this problem in more detail in the attached screencast: https://jmp.sh/7nx0vx51

Please add Kyrgyz language

Kyrgyz voice input is currently available only in ChatGPT. Other major companies, including Xiaomi, Google, Apple, Windows, and Linux, still do not provide full support for Kyrgyz voice input. In Kyrgyzstan, a country of more than 7.5 million people, there is a genuine need for this feature. At present, many of us are forced to use Russian for voice input. We strongly urge developers to add support for the Kyrgyz language. In today’s world, a language is much more than a means of communication — it is an essential part of a nation’s culture, identity, and informational independence. Recent events, including the situation in Ukraine, have demonstrated how important and sensitive the issue of language and its role in society can be. If Kyrgyz is not actively developed and fully supported by modern digital technologies, Russian will continue to dominate the digital environment, while Kyrgyz may gradually be pushed aside. Supporting Kyrgyz voice input is therefore not merely a matter of convenience — it is an important contribution to preserving, developing, and securing the future of our language.

When will you add Kyrgyz Language? (https://ru.wikipedia.org/wiki/Кыргызстан)

Kyrgyz voice input is currently available only in ChatGPT. Other major companies, such as Xiaomi, Google, Apple, Windows, and Linux, still do not offer full support for Kyrgyz voice input. In Kyrgyzstan, a country with a population of more than 7.5 million people, there is a real need for this feature. At the moment, many of us have to use Russian for voice input.

Please can we chnage the order of modes in the list. If you have several "button" that are relating to a language (such as informal and formal text) it would be great to move them next to each other

Slightly annoying forced tutorial

This is kinda annoying actually, if you already use this and just want to log into another machine, it keeps forcing you to do the tutorial and it seems that if you just close it, it does not register you user and you have to log in everyday again until you do the very exciting tutorial … seems a matter of just adding a ‘skip tutorial’ feature, hope it is.

Interactive Toolbar

I use three monitors. Is it possible to select which monitor has the interactive toolbar on it or even better, is it possible to have it on all three?

Adding the GPT-5.3 model to custom modes

The yesterday a new model, GPT-5.2, was released. I’d really like to see this model in the list of options for custom mods. But if you do add it, please make sure to add the non-reasoning version, because it takes a very long time to think, and overall the cost of the reasoning model for you will be about 5–7 times higher than the non-reasoning one. That’s just not economically reasonable, and for the user, such long reasoning isn’t really desirable either.

i am expecting android version of Blabby AI from you , Thank you

Customize the position of the floating bar.

I would like to be able to move the bar on the Windows app just like I am able to move the dot or bar in the browser add-on. To put it wherever I need it, maybe on the second monitor.

Beta Version of Windows: First Impressions and a Wish for More Flexibility

I just tried out the beta version of Windows, and overall, I’m positively surprised. The system already makes a good impression. However, I would have liked more flexibility with the taskbar. Since I work with multiple screens, it would be ideal if I could freely move the taskbar between monitors. A feature that allows the taskbar to dynamically adapt to the desired screen would significantly improve usability.

Please incorporate a Speech Timeout feature.

Visual Studio Code Speech Extension example. it's really helpful.

Update interrupts transcription

The update notification is shown after ending a recording, interrupting the transcription. Nothing is saved to history.

Too many failed transcriptions

Something wrong happening today? Transcriptions all stopped working and it seems I still have a lot of credit left and am using the latest version

Can't handle pauses

I find that if I hold down the shortcut and maybe don't start to talk immediately or maybe pause after starting, it goofs it up and I get gibberish. I also have Wispr Flow and I don't experience anything like that with it, so it seems like something should be able to be done.

History +Wispr flow

I think it's best to implement all the best features from Whisper Flow, add history, and other useful tools like formalizing writing and breaking it into paragraphs and so on. Perhaps include options for different speeds and precision, although everything is already quite good, fast, and faster than Whisper Flow. Maybe consider some dynamic pricing, like charging based on usage volume instead of a subscription. And let it show notifications, like whisper typing, about how much time is saved monthly, annually, or in total.

Timeout too long in frequent lags

Folks, I love this tool, byt lately It is being quite annoying if I am honest … having it take what seems more than a minute to timeout when it lags … and that has been happening more often than I could just ignore it … for instance, now, perhaps 2 minutes after it started lagging, I chose to write the sentence myself at this point … and out of nowhere, as I am writing something completely different in another environment, it just pastes the content that I had dictated minutes ago …. and I also notice that for some reason it logged me out … things that are bearable when they happen like once a week, but kinda gets in the way of getting things done when they start to happen more than once a day (3rd time today already) … honestly I would prefere not to have it so slow of course, but if it is unavoidable as it seems to be, I wish I could at least set the timeout to be something shorter or just cancel with a button so at least it does not paste unappropriate content in random input boxes when I already forgot what I was trying to do. Really hoping this is fixable

Adding a rewriting feature

Are there any plans to add a rewriting feature to the Windows app in June 2025—the same one that was introduced for the browser extension?

✅Fixed - Service issues update

EDIT: Hello everyone, we have successfully migrated away from the affected area. If you didn’t receive the update in the Windows app, please download the latest version (1.5.3). Hi everyone, We’re currently experiencing issues with one of our third party providers, affecting parts of the service. This appears to be related to regional instability due to the ongoing conflict between Israel and Iran, which has been impacting infrastructure and cloud services in the region. (Reuters) We’re actively working on moving our services out of the affected regions and monitoring the situation closely. Right now we’re running in safe mode. Dictation should still work (even if it shows 0 credits), but some other features may be limited or not working properly. Affected features Login system Custom Modes (while logged in): Cannot create modes Cannot update modes Cannot delete modes Workaround If you need to create, edit, or delete modes, you can log out and perform those actions. Thanks for your patience. We’ll update you here once everything is fully stable again. If you want to stay updated, please subscribe to this post. — BlabbyAI

make it more intelligent by default

Basically in whisper flow whenever you list something it automatically structure the prompt. For example, if I'm saying something and it contains multiple options, it will automatically convert it into the list which blebby AI doesn't do. It would be a cool feature to have.

New Models needed

Hey, some feedback on the model side of things. AI moves fast, and the models Blabby is using for the second "mode" layer are starting to feel dated. There are newer, cheaper, faster, smarter options out now. To be clear, I still like Blabby better, but even Codex just added transcription to their desktop app running the newer stuff, with global hotkey activation, so the gap is closing. For the AI modes/post-processing layer, gpt-5.6-luna as the fast default and gpt-5.6-terra as the higher-quality option, plus Gemini 3.6 Flash and 3.5 Flash-Lite. Honestly, I'd love to see Gemini 3.6 Flash as it just came out and it's extremely token-efficient. For the initial transcription layer, nothing really beats Whisper V3 Large on Groq for speed and cost (~$0.11/hr at waaay past real-time). However, for a more premium option/tier, ElevenLabs Scribe v2 makes roughly half the errors Whisper does. It is 3x the price, but I think plenty of users would happily pay a bit extra for the increased accuracy. Would love to see Blabby stay ahead here. Thanks for the great service!

AI interpretation.

Sometimes it takes a text and treats it like a prompt to an AI. So if I say some sentences, it then translates them into German whenever the word “German” is mentioned in the text. Sometimes it goes off on its own and does a full analysis based on the text. For example, if I say the Earth is flat, it launches into a long explanation of why I’m wrong.

clipboard mixups.

I have been using the desktop version. The Chrome version has always worked well. I have been finding on the desktop version that frequently Blabby is not inserting the dictated text. Instead, it is inserting the last thing that was on the Windows clipboard. If I check on the toolbar, I can see that the dictation has been captured on the most recent list. However, it's not inserting when I use it. This is happening in multiple programs including Word and Firefox. Often if I completely restart the program, it will start working again. I think the behavior starts when I copy something to the clipboard in Windows.

Remove Admin Privileges

Refactor technologies that currently require elevated privileges. This will allow the application to be launched without the need for administrator rights, eliminating the hassle of entering passwords or dealing with permission prompts. Improves convenience and provides a smoother, faster user experience

🛡️ Code Signing

We’re currently working on obtaining a code-signing certificate for our Windows app. Once completed, this will remove the “unknown publisher” warning and ensure a more seamless installation experience. Thank you for your patience, this is a priority for us, and we’ll update you as soon as it’s done! 🙌

Adding the GPT-5.2 model to custom modes

The day before yesterday a new model, GPT-5.2, was released. I’d really like to see this model in the list of options for custom mods. But if you do add it, please make sure to add the non-reasoning version, because it takes a very long time to think, and overall the cost of the reasoning model for you will be about 5–7 times higher than the non-reasoning one. That’s just not economically reasonable, and for the user, such long reasoning isn’t really desirable either.

Adding the GPT-5 model for mods

I was wondering if you’re planning to add GPT-5 as the default option for handling user mods in your program. I’ve been using GPT-5 myself for a while now, and it deals with complex prompts way more efficiently than GPT-4.1.

Missing Floating Toolbar

I’m testing this on Windows, and I’ve run into an issue. The floating toolbar is no longer visible. All I can see is the tray icon. When I double-click the icon, it only opens the settings window. If I minimize the settings window, the icon just sits in the taskbar, and clicking it again only brings the settings back up. If I close the settings, the icon minimizes to the Windows tray, but re-opening it still only gives me the settings—never the actual floating toolbar with the dictation modes. This wasn’t the case when I first installed the software—the toolbar was available then—but after restarting my machine, it has disappeared.