If you want to check out SuperWhisper I'd be so grateful if you use my link to get the app: superwhisper.com?via=robert Get the Alfred Workflow: github.com/ognistik/alfred-superwhisper
I love your enthusiasm - I bought the lifetime version of the app after watching your earlier videos. I'm excited to see your future videos on how to get the most out of SuperWhisper - the iPhone version is crazy good as well.
Hey, thanks for your comment! I'm glad to hear that my video inspired you to get SuperWhisper. Actually, I had some reservations about the iPhone app because I was experiencing some annoying bugs. However, just today there was an update, and the new version introduces the Ultra Cloud model for iOS along with several other improvements. It's much more reliable now, and I'm really happy with it. Planning to explore it some more!
@@hugoschmitt1772 Yeah, the iOS version of the app can be good for really long texts, and I still use it for some processing with AI, like helping me improve messages that I share online (like this reply!). I think I've even used it to outline some video ideas. I wasn't using it much before because it was buggier, but I'm happy to see it more reliable now. I'd love if it we could integrate it with iOS Shortcuts, but who knows if that'll ever happen.
Great video, Robert! Upgrading my superwhisper to a lifetime license was possibly one of the best purchases I've ever made for my MacBook! Since I do so much work in a busy/noisy areas, I spent some time looking for the best microphone that I could pair with SuperWhisper to get a clean transcription even if the room is pretty noisy. After trying out a few, I quickly found what I believe to be the best right now-the DJI Mic 2. It has a realtime voice isolation/noise cancelling mode and is absolutely incredible at blocking out background noise, even in some of the most noisy places like coffee shops. The wireless transmitter has basically zero latency and overall the DJI Mic 2 kit has taken my superwhisper use to an absolutely new level (plus there are other great features like gain staging, automatic charging case for the transmitters, offline recording memory for the transmitters etc.) Neil is a great developer and I can't wait to see what other great features get added in the future! -Cheers, Brad
I'm interested to know if you have experimented with any sort of prompt engineering for the AI modes in terms of trying to get the best/most efficient outputs. With some of the new updates adding so many features/new modes (like Super), I have been trying to find the best model combinations and prompts that have just the right amount of context to get a fast and accurate output. Some of my most used modes are relatively simple, but I've really wanted to try some heftier modes with some of the latest models like o1 to try to get the most out of them for some more advanced tasks!
One last thought! If you are looking for a really high quality and streamlined app that integrates with SuperWhisper, I would highly encourage you to check out Raycast! It is an alternative to Alfred that I've been using for the past couple months, and the speed of development in their community is really amazing! While it was hard to let go of some of my Alfred workflows, Raycast has some unique functionality, features, and extensions that are simply unavailable in Alfred. Plus there are official extensions for hundreds of apps, including Superwhisper, that are published by community developers and are updated automatically (I haven't found nearly the same amount problems with raycast extensions maintaining functionality across application updates, which is something that I did experience quite often with my Alfred workflows-I think that the Raycast community being a lot newer and more active really helps to keep the extensions up to date). While I wish there was a lifetime option for Raycast, like there is for Super Whisper or Alfred, I found that the subscription is worth it for me for how much functionality I get out of it and the overall ease of use that it offers. The Raycast RUclips channel (www.youtube.com/@raycastapp) has a bunch of amazing videos on all sorts of different features of the application. And I would definitely encourage you to check it out--I think your creativity in developing integrations could really fit well into the Raycast community!
@@bradpaul1699 Hey Brad, thanks for the tip on the noise-canceling mic! The Bluetooth mic I'm currently using (Hollyland Lark M2) is also supposed to have good noise cancellation, but I haven't tested it thoroughly-especially not in extreme conditions. As for prompting, I usually start with the prompt builder I shared in the video, making adjustments until I consistently get what I need. Sometimes, if I need something super simple I'll just skip that and give it a shot myself. One thing to note is that with the most recent updates, all built-in modes in SuperWhisper offer less customization. I think this change happened because many users had trouble with prompting, which probably was leading to unnecessary bug reports. So, currently the built-in modes (including Super) are a bit limiting, but they are definitely more user friendly and probably better for those that don't need the customization. I prefer to steer clear of all built-in and preset modes since I like to control what goes in and out. That's why I always use "Custom" for all my modes (aside from one Voice-only mode). One thing I've started doing more of lately is using XML tags, which has improved the overall results I get from custom prompts. Specifying tags like , , , and similar ones really makes a difference.
Hello Robert.I really admire your consistency in creating content. Each of your updates brings valuable insights that I learn so much from. Thank you for sharing and have a good trip in China.
Unfortunately, the best free voice-to-text app I knew was Aiko, but it seems to have switched to a paid model. I haven't looked into it since, I'm not sure about any other alternatives that can match it (I know SuperWhisper actually has a free version, but I don't know its limitations). As for text-to-voice, I honestly haven't thought of many cases so haven't really tested a lot, but I’ve heard great things about the new Kokoro model, but haven't really played with it.
Great vid Robert your enthusiasm for SW is inspiring please do more videos on how you use SW. Also you are so right Neil is a super guy you just feel though at some point he might bail because of his work load and really the sustainability of the app cost wise ..let’s hope this never happens 👍
It's true; I’m also a bit concerned about the long-term sustainability. Especially with the lifetime option, I think it currently is kind of too good to be true, so not sure how or when will this eventually change. You're also right about the workload. I see that Neil is a really nice guy, and I don't think he'd want to disappoint current users (so I don't see him turning his back on the app, at least not intentionally), but it would still build more trust if he could assemble a team of developers. It looks like he's already gotten someone helping him with support, so maybe he’s considering more options like this. I see so much potential here, and I just hope the dev can build the necessary structure to support growth and improvements.
Hi Robert. I really love your videos. Because if one of them, I have gotten Alfred, now downloaded your SuperWhisper Workflow for Alfred, but I am stuck with some specifics. I want to have specific modes output transcripts placed into a specific folder on google drive desktop. In Superwhisper so far I can only have one Folder Link, where all the transcripts are being placed. Any idea how Alfred could help with this?
Currently, SuperWhisper outputs the results to the JSON file in the recordings folder. This is what is displayed in the recording window. I previously had some keyboard maestro macros that would read the JSON file right after it was saved, so it allowed the user to automatically send the results anywhere he/she wanted. Unfortunately, SuperWhisper's recent updates changed the way in which the saving of the json file happens, and because of that there's no easy way to automate something with the results (not with Alfred or KM). The developer has mentioned there will be an official way to automate things like this. I hope this will be released in the near future. As soon as it's out I'll share about it. The only one way I can think it is possible to automate something like this right now, would be by using something like Keyboard Maestro which can recognize images... so it could trigger once the SuperWhisper menu bar icon has the green checkmark. But this method is definitely more complicated and not very reliable.
@@afadingthought Thank you so much for taking the time to answer my question Robert. So kind of you. Would you recommend another voice transcript tool, that is similar to Superwhisper and can be solving my issue with the saving of different modes?
@@Inés-x8u8u If you're looking to transcribe audio files into text files, you might want to check out MacWhisper. I use it quite a bit for creating subtitles. It also offers dictation and AI processing options which I don't find it as intuitive or convenient as Super Whisper, but it does let you save files, which Super Whisper does not.
@@robertotriunfo9674 Yes. There are plans for an "officially supported" way to hook automations to SuperWhisper. I'm just waiting for that feature to be released and will explain this and more.
Ultra Cloud is amazing. There were a few server issues with this model in the last few days, but everything seems to have improved now. The one I use for backup as a local voice model is Ultra Turbo V3. Pretty good, just a bit slower.
If you want to check out SuperWhisper I'd be so grateful if you use my link to get the app: superwhisper.com?via=robert
Get the Alfred Workflow:
github.com/ognistik/alfred-superwhisper
I love your enthusiasm - I bought the lifetime version of the app after watching your earlier videos. I'm excited to see your future videos on how to get the most out of SuperWhisper - the iPhone version is crazy good as well.
Have you seen the previous one where controls his whole computer with dictation and a few Alfred workflows? Amazing stuff.
Hey, thanks for your comment! I'm glad to hear that my video inspired you to get SuperWhisper. Actually, I had some reservations about the iPhone app because I was experiencing some annoying bugs. However, just today there was an update, and the new version introduces the Ultra Cloud model for iOS along with several other improvements. It's much more reliable now, and I'm really happy with it. Planning to explore it some more!
@@afadingthought Yeah I couldn't understand the "workflow" with iOS. And there, native dictation works so well... Maybe for really long texts?
@@hugoschmitt1772 Yeah, the iOS version of the app can be good for really long texts, and I still use it for some processing with AI, like helping me improve messages that I share online (like this reply!). I think I've even used it to outline some video ideas. I wasn't using it much before because it was buggier, but I'm happy to see it more reliable now. I'd love if it we could integrate it with iOS Shortcuts, but who knows if that'll ever happen.
Great video, Robert! Upgrading my superwhisper to a lifetime license was possibly one of the best purchases I've ever made for my MacBook! Since I do so much work in a busy/noisy areas, I spent some time looking for the best microphone that I could pair with SuperWhisper to get a clean transcription even if the room is pretty noisy. After trying out a few, I quickly found what I believe to be the best right now-the DJI Mic 2. It has a realtime voice isolation/noise cancelling mode and is absolutely incredible at blocking out background noise, even in some of the most noisy places like coffee shops. The wireless transmitter has basically zero latency and overall the DJI Mic 2 kit has taken my superwhisper use to an absolutely new level (plus there are other great features like gain staging, automatic charging case for the transmitters, offline recording memory for the transmitters etc.) Neil is a great developer and I can't wait to see what other great features get added in the future! -Cheers, Brad
I'm interested to know if you have experimented with any sort of prompt engineering for the AI modes in terms of trying to get the best/most efficient outputs. With some of the new updates adding so many features/new modes (like Super), I have been trying to find the best model combinations and prompts that have just the right amount of context to get a fast and accurate output. Some of my most used modes are relatively simple, but I've really wanted to try some heftier modes with some of the latest models like o1 to try to get the most out of them for some more advanced tasks!
One last thought! If you are looking for a really high quality and streamlined app that integrates with SuperWhisper, I would highly encourage you to check out Raycast! It is an alternative to Alfred that I've been using for the past couple months, and the speed of development in their community is really amazing! While it was hard to let go of some of my Alfred workflows, Raycast has some unique functionality, features, and extensions that are simply unavailable in Alfred. Plus there are official extensions for hundreds of apps, including Superwhisper, that are published by community developers and are updated automatically (I haven't found nearly the same amount problems with raycast extensions maintaining functionality across application updates, which is something that I did experience quite often with my Alfred workflows-I think that the Raycast community being a lot newer and more active really helps to keep the extensions up to date). While I wish there was a lifetime option for Raycast, like there is for Super Whisper or Alfred, I found that the subscription is worth it for me for how much functionality I get out of it and the overall ease of use that it offers. The Raycast RUclips channel (www.youtube.com/@raycastapp) has a bunch of amazing videos on all sorts of different features of the application. And I would definitely encourage you to check it out--I think your creativity in developing integrations could really fit well into the Raycast community!
@@bradpaul1699 Hey Brad, thanks for the tip on the noise-canceling mic! The Bluetooth mic I'm currently using (Hollyland Lark M2) is also supposed to have good noise cancellation, but I haven't tested it thoroughly-especially not in extreme conditions.
As for prompting, I usually start with the prompt builder I shared in the video, making adjustments until I consistently get what I need. Sometimes, if I need something super simple I'll just skip that and give it a shot myself. One thing to note is that with the most recent updates, all built-in modes in SuperWhisper offer less customization. I think this change happened because many users had trouble with prompting, which probably was leading to unnecessary bug reports. So, currently the built-in modes (including Super) are a bit limiting, but they are definitely more user friendly and probably better for those that don't need the customization. I prefer to steer clear of all built-in and preset modes since I like to control what goes in and out. That's why I always use "Custom" for all my modes (aside from one Voice-only mode).
One thing I've started doing more of lately is using XML tags, which has improved the overall results I get from custom prompts. Specifying tags like , , , and similar ones really makes a difference.
Hello Robert.I really admire your consistency in creating content. Each of your updates brings valuable insights that I learn so much from. Thank you for sharing and have a good trip in China.
Thank you. Which good and free AI-tools are out there who can voice-to-text and also text-to-voice?
Unfortunately, the best free voice-to-text app I knew was Aiko, but it seems to have switched to a paid model. I haven't looked into it since, I'm not sure about any other alternatives that can match it (I know SuperWhisper actually has a free version, but I don't know its limitations). As for text-to-voice, I honestly haven't thought of many cases so haven't really tested a lot, but I’ve heard great things about the new Kokoro model, but haven't really played with it.
@@afadingthought Thanks for your reply! :)
Great vid Robert your enthusiasm for SW is inspiring please do more videos on how you use SW. Also you are so right Neil is a super guy you just feel though at some point he might bail because of his work load and really the sustainability of the app cost wise ..let’s hope this never happens 👍
It's true; I’m also a bit concerned about the long-term sustainability. Especially with the lifetime option, I think it currently is kind of too good to be true, so not sure how or when will this eventually change. You're also right about the workload. I see that Neil is a really nice guy, and I don't think he'd want to disappoint current users (so I don't see him turning his back on the app, at least not intentionally), but it would still build more trust if he could assemble a team of developers. It looks like he's already gotten someone helping him with support, so maybe he’s considering more options like this. I see so much potential here, and I just hope the dev can build the necessary structure to support growth and improvements.
Hi Robert. I really love your videos. Because if one of them, I have gotten Alfred, now downloaded your SuperWhisper Workflow for Alfred, but I am stuck with some specifics. I want to have specific modes output transcripts placed into a specific folder on google drive desktop. In Superwhisper so far I can only have one Folder Link, where all the transcripts are being placed. Any idea how Alfred could help with this?
Currently, SuperWhisper outputs the results to the JSON file in the recordings folder. This is what is displayed in the recording window. I previously had some keyboard maestro macros that would read the JSON file right after it was saved, so it allowed the user to automatically send the results anywhere he/she wanted. Unfortunately, SuperWhisper's recent updates changed the way in which the saving of the json file happens, and because of that there's no easy way to automate something with the results (not with Alfred or KM). The developer has mentioned there will be an official way to automate things like this. I hope this will be released in the near future. As soon as it's out I'll share about it.
The only one way I can think it is possible to automate something like this right now, would be by using something like Keyboard Maestro which can recognize images... so it could trigger once the SuperWhisper menu bar icon has the green checkmark. But this method is definitely more complicated and not very reliable.
@@afadingthought Thank you so much for taking the time to answer my question Robert. So kind of you. Would you recommend another voice transcript tool, that is similar to Superwhisper and can be solving my issue with the saving of different modes?
@@Inés-x8u8u If you're looking to transcribe audio files into text files, you might want to check out MacWhisper. I use it quite a bit for creating subtitles. It also offers dictation and AI processing options which I don't find it as intuitive or convenient as Super Whisper, but it does let you save files, which Super Whisper does not.
Do you have plans to show how you hooked up superwhisper to your notes automation?
@@robertotriunfo9674 Yes. There are plans for an "officially supported" way to hook automations to SuperWhisper. I'm just waiting for that feature to be released and will explain this and more.
I love SuperWhisper! But yes I feel like I am not getting full use out of it so I am looking forward to your future videos :)
I really like your videos and just typed this comment using the latest Ultra model, which, for some reason, I was not using yet.
Ultra Cloud is amazing. There were a few server issues with this model in the last few days, but everything seems to have improved now. The one I use for backup as a local voice model is Ultra Turbo V3. Pretty good, just a bit slower.
Thanks for the great information. Looking furhter into if I have enough use cases i could use the app for and justify it's lifetime cost.
Buenísimo amigo