Be My Eyes just released an app for Windows. It is available in the Microsoft Store. It will let you describe your screen, the image in your camera, in a file or in the clipboard. It is very bare-bones in terms of settings, strictly speaking there are no settings, you can't even change language as of yet. The app seems to work fine otherwise, it delivers descriptions with about the same speed as the iOS app, maybe even a bit faster.
Comments
Be MY AI on Windows improvements
A. I want the window to be not visible in my alt-tab app-switcher unless I specifically call up the chat window by requesting a description or activating the app icon.
B. The window should be minimized to the notification area of Windows when user closes the app.
C. Currently along with Be My AI window, there's another blank window that shows up on my app-switcher, and closing that window causes those keyboard shortcuts to stop working. This shouldn't happen. Space in my taskbar or app-switcher is valuable, don't consume it with meaningless windows.
D. Give me a good, easy to understand privacy policy.
E. Give ability to customize those keyboard shortcuts.
F. Add Video support. This might be the most important request yet.
video support
I heard an interview with Mike Buckley, the CE/O of Be My Eyes on the Double Tap podcast a few days ago. Apparently video descriptions are coming when video supportfor GPT 4?O is released, but he has no idea when this will be. And yes, a clear privacy policy will put more minds at ease.
it was a cat playing Rachmaninov
Hi Lottie,
Even though I don't watch videos about cats, I've had frames from videos described to me which are pretty accurate. I have no reason to doubt the accuracy of the description based on what I know about the content creators. Oh, it was a cat all right. It's fascinating to finally learn what visual aids people use in their videos, and how it all fits together with the spoken dialogue. This is all so exciting. I think I'm starting to finally understand why a song might get popular just because of the video. I still don't agree with people's take though. A song should be about the music! But I understand it from a theoretical point of view.
Lottie
Cat playing music? I am curious if I can get she who should not be name to do ACDC or Scorpion?
be my eyes and describing windows
If I'm able to get it to describe a still from a video, how does one have be my eyes describe the screen?
Describing the screen
Hi Michael,
Make sure Be My Eyes is open, go back to the video you're watching, and then press ALT + CTRL + H. If you wait a couple of seconds you'll get a description. You could potentially keep doing this throughout a whole video because the descriptions come back so quickly. You can Alt + Tab to read the contents of the Be My Eyes window if you want to catch up with everything, but it reads everything automatically anyway. Make sure your browser window is maximised, having your browser in fullscreen might help too.
yeah but...
You have to do it frame by frame, which is tedious, it's awesome that it can do it and one day we'll be able to get it to describe things like an audio describer, I'm sure of it, but for now this is as far as it goes.
How does one
Get youtube videos described using picture smart that is? sorry for being a cave-dweller... but is there any specific keystroke that one uses to get a frame-by-frame describtion on a continuous manor?
window description
I find the cmd ctrl alt h hit and miss even with it being open.
this is a great start but they have a lot of work refining it. I definitely agree with the previous poster where there should be a way to minamize it to the system tray at the very least.
cmd = command, right?
It should just be alt control h.
Or were you writing cmd for those using a mac who need to go over to windows?
@Gokul
I don't think you can, you have to do it manualy.
BME and X
Formally twitter. People may or may not know that X takes over the keyboard or certainly it does with JAWS so this means that trying to use BME to describe images doesn't work. Unless, someone has found a way of either copying them which seems impossible or another way.
Sorry not very clear
Sorry that wasn't clear. You can use CNTL+ALT+h but BME reads a shed load of stuff including the tweet etc. To get a description of the picture/image is the issue. have'nt found a way of exporting it or anything else for that matter.
yeah, it'll do that.
It's annoying but at the moment there's not much you can do apart from write to them.
It's a bit to verbose and I have to tell it to cut out all the fluf and it does for a couple of times but I don't think it remembers what you say when you close the app.
picture smart still appearing to be a better option
Based on my testing, I still prefer picture smart from jaws. One major cause for this is describing a picture in file explorer. In be my eyes, one has to browse to the file then have it described whereas with picture smart, it's just a command when you have the photo selected in file explorer.
I'm also finding a lot of inconsistencies with having be my eyes describe a still on youtube. For instance, I tried doing this last night and rather than describing the still, be my eyes began describing all my opened tabs and windows, this is despite the video being in full screen and maximized.
Picture smart
Yesterday I was watching a old nice TV show. Kate and Allie. I use it to describe a seenary and it gave me a good description of Kate yellow shirt and Allie blue shirt. Told me who was dark hair and who was a blond and type of blond. Nice job JAWS.
hopefully real time description in the future
I hope with the GPT 4 O... it can have real time discribtion
in the future
Finally used it
I finally installed the Be My Eyes app yesterday for Windows. It actually worked really well. I had an option in Wordpress, for a plugin that does not use standard controls, and couldn't tell if said option was checked. After getting a description of what was on screen, I asked if said option was checked, it said no, hit enter on it, had it check the screen again, asked again, and it was that time, so really helped a lot.
Be My Eyes and Picture Smart
So, at first, I was going to suggest that compairing them is not fair, one is screen reader, another is just an app. But then I thought, Be My Eyes is an app, a dedicated app for doing 1 thing, describing things.
This is not that hard to achieve considering apps like Quicklook from microsoft store have successfully implemented quickly showing the selected file preview on scree if you press certain keyboard shortcut.
NVDA's Advance OCR, NAO addon is able to interprit the selected file from file explorer without opening the file in any other app.
So I am guessing this shouldn't be a hard functionality to implement. The concern is that should Be My Eyes have access to files and folders at all? I mean grabbing screenshots on command is one thing, but uploading files from my local storage to an AI on the clowd. This might quickly get out of hand. Further, if I end up selecting some 4K wallpaper, it's going to waist time uploading a very hevy file to the clowd, whereas slightly lower resolution might have surved the purpose. Compression based on file size might have to be implemented.
Given Be My Eyes's privacy policy and T&C currently, I would rather not let it handle my files for now.
They siriously need to get their game figured out around user privacy. Open AI or any other company shouldn't be using our images of friends and family, our documents to train it's AI, and, sell that AI for proffit later for god knows what purposes.
If you are performing a good will service, be transparent. If you have shaidy things in mind, remain scylent. We will get the message.
That makes me wonder, how on earth copyright and things like that would work? Currently, lot of concerns are being raised around how Open AI trained it's models, without having any copyright agreements with the content provider or creator. Now, say Youtube doesn't have the deal with Open AI, but I use Be My Eyes to get some video seens described, am I violating some terms and conditions? Am I going to need some Marrakesh 2.0 Treaty to sort this out?
I don't care to much.
I don't really want my data to be solled to other companies but if they want to train their AIs, go for it.
BeMyEyes has updated and added a couple of sounds, you can turn all of them off in the settings tab and it could just be me but I think it's gotten a little bit smarter too.
Just wanted to say how great this ap is!
I use BME on Windows a lot. One of the things I like most is being able to copy images to the clipbaord and get them described - this is great on Facebook, where the built-in alt text is rubbish.
Agreed
While I cannot speak for Facebook, as I do not use it, Be My Eyes is fabulous on PC.
True story~ 😀
It really is.
When we get access to GPT 4.0 I think that will be another game changer.
Oops, I only remembered just now
they own this site now. I was genuinely feeling happy about it as I was using it!
I wonder when we will get…
I wonder when we will get the Mac version. Would be very nice to have.
Needs work, but very useful
I've been using this quite a lot since it was released. I'm an avid gamer, and sometimes NVDA can't read certain UI elements or text that doesn't have good contrast. It's also really helpful in describing scenes. I can't wait till the video sharing feature becomes reality. I would like to be able to set some custom prompts before taking the picture however. I want this for the mobile app and the Windows app, because you really have to coax information out of it sometimes. Still, it is a literal game changer, and a huge foreshadowing of what is possible in the future.
Smart picture and facebook
Smart picture works well and does a great job on facebook. Describe well and I can use other AI to get more info or ask questions. Works on videos also. Be my AI on the iPhone is OK but does not do a good job.Probably the same with the windows app.
Is Smart Picture a thing?
Or are you talking about the JAWS feature with a similar name? I know FB are bringing new things out all the time...
Also, in what way does BeMyEyes not work wel on your iPhone?
Mac Version
A Mac version would be great. I used the windows version and it was a good experience. Be MY AI should come to mac. :)
I am so glad!
Usually it's the other way around, new fancy applications launching on Apple side first, and later coming to Android and Windows. The Be My Eyes team has truly recognized where the target audience is, and accordingly delivered. Hats off for that.
Hope Mac users get to enjoy this amazing capability soon.
Isn't there a ChatGPT or OPEN AI app for the Mac?
If so, is it useable and can it do the job of BME? I think I have always assumed that this was why there was no BME for Mac.
GPT for Mac
Yes Lottie,
ChatGPT is available for the mac.
https://openai.com/chatgpt/mac/
You can share your screen and do all sorts with it. From what others have said it's accessible with VoiceOver on the mac too. According to the above page it's not still not available for Windows yet..
Charlotte Joanne
Does not provide detail info or tells it can not do so. JAWS smart picture does so. Be my AI sometimes say things is not true. Picture of a friend she uses an Umbrella, in JAWS she does not, I ask her when the picture was taking and she did not had one.
Wow, Holger , that's not good
On the phone, i use it mostly for images from Mastodon. It has neverd refused and it seems pretty good.
Of course, I'm not defusing bombs, so I don't double check that often. But still, when I do, it seems v good and it is great at explaining funny cartoons!
JAWS and picture smart
Picture smart uses GPT and Claude I think? So you might get better or different results sometimes.
There's a Mistral model with vision now
You can use it, for free, via Le Chat. It seems to give the same sort of descriptions as the rest.
JAWS and Picture Smart
When using Picture Smart, when holding down the shift key with any of the Picture Smart layered hotkeys, JAWS will provide feedback from both Gemini and Claude. "this provides having the feedback from both services serves as a check on how well the AI has responded. You can also go on to ask additional clarifying questions. FS did a really nice job with this feature.
--Pete