From Your Eyes

While I could start by saying the focus in the app is garbage and its impossible to choose a country code for the phone number, I'll focus on the descriptions. I'm not sure what AI model this is running against, but the descriptions are useless and garbage. I'm not even sure it is running against AI, as the descriptions came back in butchered english. All that said, I would never trust an app that doesn't have a description in the app store in english. Sorry, but seems ultra sketchy to me.

I can’t even get past the initial registration stage not sure what language it is in but certainly not English

Perhaps the developer should’ve consulted with the Blind community about the accessibility of their app? I mean, this is basic blindness 101…

Also, I don’t so much have a problem with the broken English as the fact that they’re asking for a bunch of personal information without giving an explanation as to why they need it.

I had real hopes for this app as there was one like it or similar to it in the past that got defunct, but I quickly deleted this app off of my phone after my own that experience trying to sign in, my suspicions and reading the comments on this forum.

I'm not downloading it do to the reason everyone else isn't. Here's my oppinion for what it's worth if you are going to make an app available in the united states put the thing in english, even the youtube demos aren't in english.

I just checked out a couple of screens of the sign up process and the app works fine.

Hi, I have downloaded it - I am from Italy. And the app is in turkish; there is a button at the start which is called "en", and that switches the app in English. Interface, I mean.
I wrote to the support, they opened a mail app with info @ fromyoureyes . app
and I reported the problem with language and the country code (I have registered through the help of a sighted person).
But, the worst is that images description by AI, are unreliable! There is also the web app - https://www.fromyoureyes.app/app/app/login.html
Well, the fact it's a html web app makes me feel it's projected with old technologies, but, letting that go, descriptions are well done but not corresponding to the reality!
I mean: you can give me the most suggestive description of the world, but if you describe an elephant while it's a turtle..
and, the web app has no access to registration, and no access to human descriptions.

this is actually a very nice app. there was an app called "bespecular", this one is just like that. it's funny some people loved bespecular but think this one is trash.

the ai is hit and miss. and when it works, it works very well, too. here is an example:

"In this image, there are three light gray-colored cups, one at the top, one at the bottom right, and another at the bottom left. Two silver-colored cups are located in the center of the picture. Also, a lavender-colored book is present at the top left. It seems that this is an indoor scene. The color choice of the cups and the book is quite pleasant, suggesting a comfortable and soothing atmosphere."
it is accurate, by the way. and if the ai doesn't work, you still got human, there is no need to be so hostile to something you don't at least invest some minuts to see how it works.

When I tap and add the photo to the app for it to give me a description it doesn’t do anything, I’m not sure what I’m actually doing wrong it just says photo loading. I have tried adding a photo from the gallery but it makes no difference.

It’s a long wait 😂 it’s all good, I’ve taken the app off but worth a go anyway

Don't think she's the developer herself but it was a friend of mine who came up with this idea and then got funded by certain organizations/companies to release the app. I've personally heard statistics from her on how From Your Eyes compared to other similar apps and services. So she'd say both text recognition and image descriptions provided by From Your Eyes' AI would be more accurate than other counterparts. I doubt I'll be able to find anything in English on that, and don't have the necessary means to verify her claims myself, but I can definitely say some have been too harsh on this app. She's also been talking about Chat GPT integration and beta testing is already under way. This app, unlike Be My Eyes, is also expected to describe videos. I don't and likely won't ever mean to defend her or act as her voluntary advertiser, as we already have strained relations for a while, but I strongly recommend you to just calm down and put your prejudice aside. This friend can also communicate in English herself without any major issues, and has also been bragging about how users from abroad have been making up a higher portion of the total users. So I don't think she'll deliberately ignore any feedback. I don't really think I've ever encountered such harsh remarks on apps lacking support for English released by developers from countries like China or Russia, and even Europeans. From Your Eyes, on the other hand, is being treated as an app with no support for English whatsoever, even though this is not at all the case. Some have been reporting bugs even though I don't remember having some of them myself, but do bear in mind that I am not the app's developer, nor am I a representative of the company.

Firstly, I got the app interface in fine English. Seems the developer has worked on the language issues. However, the country-code selection bug during the sign-up can really be a show-stopper. I could make it work after multiple attempts.

Coming to the description part, I found the description mostly accurate and detailed. At times I encountered completely inaccurate description, though.

The gallery uploads appear to be working well, though they are taking some time to be processed. But I could not get the upload from camera processed even after half an hour wait!

Overall, though, I would say that the app looks promising.

Why is information like my e-mail address and phone number required for this app?

I take back what I said earlier. I did download and with the improvements I feel more comfortible about using this app. It looks promising!

Hello. I'm Zülal, the founder and CEO of FROM YOUR EYES. I'm also one of the developers and visually impaired myself. I have read all your valuable comments and would like to provide a collective response to address your concerns and facilitate direct communication with me.
The issues mentioned in the comments have been resolved through performance improvements. Here are some usage tips:

1. Staying on the loading screen after signing up for the app: This issue has been completely resolved. After signing up for the app, you will be directly redirected to the home page.

2. App languages: Our app currently supports service in fifteen of the most commonly used languages. We expect you to select either English or Turkish as an option on the initial screen to proceed. Once you log into your account, you can choose your preferred language from the "Language Preferences" tab in the "Preferences" section. We are working to make this process more user-friendly from the moment the app is launched. In the next update, the app's language will be determined based on your device's language when you directly download the app.

3. Integration of FROM YOUR EYES and GPT: The artificial intelligence of FROM YOUR EYES can currently recognize 865 colors, perform OCR with an accuracy above 85%, and identify images with a 94% accuracy rate. These rates exceed the standards. Through the integration with the GPT language model, the descriptions provided by FROM YOUR EYES are narrativized before reaching you. This means that GPT and FROM YOUR EYES work together to deliver the best descriptions to you. As the number of visuals in the dataset increases, FROM YOUR EYES will continue to provide even more accurate descriptions with compelling narratives. We are continuing to train the system by adding more visuals to the dataset.

4. The main difference from Be My Eyes: FROM YOUR EYES delivers personalized descriptions to its users. We send description outputs based on each user's preferences. So even if the same image is uploaded by two different people, two different outputs will be generated. In this regard, our artificial intelligence functions as a personal assistant.

5. Role of volunteer descriptors: You can approach the volunteer descriptors of the app when you want to contribute to the development of artificial intelligence descriptions in specific areas. Simply press the "Insufficient" button at the bottom of the screen. They will improve the descriptions and send them back to you.

6. Web and mobile app together: We have launched a web application for FROM YOUR EYES on our website. We released this application so that users can access our artificial intelligence while using a computer.

7. The story and demo of FROM YOUR EYES: We continue to produce more English resources about FROM YOUR EYES. One of these resources is a detailed explanatory video. You can access it here:

https://www.youtube.com/watch?v=S7lhjulCj9A

the previous received description always get replaced and disappears. so if one receive an answer from a human descriptor, the receives a second one, the only thing remains is the second answer. so if i walk away for an hour, come back and receive 4 answers and go to check the descriptions i receive, all i can see is the last one. maybe you can adjust this.

Thank you so much. We appreciate your input and we'll definitely consider it. Your ideas and feedback help us improve and provide a better experience for our users. Thank you for your support and keep the great suggestions coming!

Yay, here's Zülal herself to promote her app and respond to your comments and feedback! Or was it "feedforward", as she would often put it? Uh, whatever. Anyway, regarding the personal information issue, the website (fromyoureyes.app) does have a link named "KVKK", which can almost be considered the equivalent of the GDPR, but clicking/activating it unfortunately doesn't bring up anything, so I don't know much about that either. Hope this also gets fixed soon, with the same terms written in English and other languages as well so that users make sure their privacy is not invaded, at least without their consent.

First, I think it is fantastic that one of the developers is on this forum! That's such a quick response to addressing many questions at one time.

Just to ask again and confirm: I believe it is Misty who has asked about the need for email/phone number upon creating an account. this is pretty common, especially for the email, of course. I would think that this information is needed to establish that the user signing up is a human? Just a thought.

Hi. I am now in the midst of signing up for an account with the app. I double-tapped on TR and then chose EN (English on the next screen. It was unclear whether my selection was actually chosen, since I didn't hear "Selected," for instance. When going to the next screen, I found the button to sign up. I have filled in most of my information and am trying to sign up. However, several things are issues right now for me.

1. After entering my phone number, and swiping right, VoiceOver says: +90, image, a red and white logo on a gray background." Could you please change this 2 Choose country code or something more descriptive?

2. Image descriptions were turned on by default, so I didn't realize that the image of +90 was something I needed to double-tap on to know that I should enter my country code. Maybe obvious to someone, but not to me. Once I turned off Image Descriptions and double-tapped on the +90 item, then, I could only find Afghanistan. Each time I double-tap the +90 button, I go to the next country in the alphabetical listing. It will take me a very, very long time to get to the United States of America. Did anyone find a workaround for this?

3. Next: The text needs to be edited here: "Please the accept the terms of use."

4. There doesn't appear to be a checkbox so I can agree to the terms of use. When I click to view the privacy policy, it is not in English. So, if I did want to read it I cannot do this within the app.

So, without being able to choose my country and agree to the privacy policy, I cannot proceed. If there's anything I'm missing, someone can happily correct me. I am looking forward to trying this app when I can sign up.

All the issues you mentioned above are the same ones I’m having.

Indeed, it would have been best had the developer tested with both native English speakers and with the blind community before releasing this app. Also, it is my belief that, if one part of the app is to be released in English, the entirety of the app should also have English, and unbroken, understandable English at that.

Just some thoughts.

I will once again ask what the personal information is for.

I am also actually a bit concerned about the “personalized” AI descriptions mentioned above. May I ask how this personalization is determined?

As far as I know, you will answer a few questions, or specify your preferences as to how you wish to receive future descriptions. For instance, you may not want to waste time with information on colors or want the AI to give specifically more detailed descriptions on one particular type of thing or attribute. Any future descriptions you get will be based on your preferences. Agree with you on the other points, though it would've been better had you posted them in one single comment.

I tried this app this morning. Well I tried to try it but was unable to sign up. Every time I tried to select +44 from country code options I was unable to get past Belgium. For a app in the assistive tech space that specifically targets blind people, well lets just say I’m surprised at the lack of voiceover support.

Hello,
It would be good if we could upload photos by using the share sheet, just like we can do with seeing ai. This way, we can use the app in more areas and we could support the improved descriptions more.
Thanks in advance.

Why worry? Big Brother will come for you whether you want him to or not.

I can't even delete my account, but that could be due to my internet speed.

I'll keep trying.

if one upload HEIC from the photo gallery, the operation will altimately fail.

Hello everyone,

First of all, we want to thank you for supporting our progress.
Every comment here is valuable to us.
Being an early-stage and rapidly evolving technology, one of the things our team excels at is growing together in our development areas.
Since the platform doesn't allow me to individually respond to each comment, I will provide a collective response.

1. FROM YOUR EYES started its journey with a beta team for visually impaired users, spanning across 18 different countries since its prototyping phase in April 2022. This allowed us to create a seamless app in terms of accessibility. Furthermore, as one of the founders and developers, my own visual impairment has positively influenced this process.

2. Request for localization support in-app languages: If you feel that understanding the app in English is difficult, you can support us in localizing it. In this regard, you can communicate with us closely. We can do this not only for English but also for the 15 languages currently included in the app. You can simply send us an email using the "Contact Us" button within the app.

3. Slowness in the speed of descriptions: Our artificial intelligence identifies images within 7 seconds, but due to some infrastructure changes, delivering them to you takes more than a minute. We will solve this issue by the end of this week. We plan for the app to be faster by the weekend.

4. Storage of account information in the app: We store account information in the app because we provide services to both visually impaired users and volunteer describers with two different account types. Additionally, in order to deliver personalized descriptions to visually impaired users, our algorithm needs to analyze their personal preferences. All data in the app's data library is stored anonymously, without any human intervention or manipulation.

5. Concerns about the language of the personal data agreement being in Turkish: You can find the same agreement on our website in English, but we will also ensure it is translated into every language within the app.

Changes in the new update:

1. The language of your app will now be determined based on your phone's language. This means that after this update, your app will be translated into your primary language preference. If you are using one of the 15 attached languages in the app, there will be no issues with your phone's language.

2. Country code shifting issue in the phone number field: We have identified a focus issue with VoiceOver. For now, the area code will be automatically selected based on your phone's language.

Aside from these, we are sure you have more questions, and we also have more explanations. At this point, we would love to come together with you to discuss these in detail. You can contact us to schedule a meeting.

We would love to get to know you individually and discuss matters related to FROM YOUR EYES.

Moreover, if you are a WhatsApp user, you can join our community for faster communication.

We sincerely thank you for your genuine interest and understanding.

Best regards,
Zülal

How do we find you on whatsapp?

Ok so I am trying to get descriptions for photos. I chose from gallary and I chose different kinds of photos. Descriptions are ok but how do I make them better? Preferences I don't understand. What does the theme actually do?

I continue to hear very negative reviews of this app. It is a shame that this had to be so poorly implemented, as it could have been a great replacement for beSpecular. Maybe someone else can have a go at actually developing a better replacement, as this one doesn’t appear to be getting much better…?

Has anyone used this app to read mail? Since be my eyes was brought up here I will say I don't use be my ai or be my eyes to read mail because of privacy.

I've used it to read mail and while it doesn't always get it, it's great for a free app.

If you want guaranteed privacy from a human-assisted app, try AIRA. These days, the minutes are, so to say, golden, so it’s geting to be a bit of a luxury item, but the agents are bound to confidentiality.

Wait, no, I'm no representative, nor am I an advertiser. I'm just writing this post soon after a meeting with the company's CEO, Zülal Tannur. The new app is already available on the App Store as FYE, and I will post it either as a separate entry or here. There are several amazing features that excite me quite a lot, like describing videos, offering personalized descriptions based on;
- Questions and instructions that you may send the AI just as in Be My Eyes
- Your preferences based on these previous interactions
- Characters that you may create and switch between, each having their own style and level of detail
I took a photo of my own room and then it described it as a busy and well-lit street having buildings on both sides. After I provided the context and stated that it should see the inside of a room in the photo, it gave the description just right. Why are you trying to narrate such a failure as a tremendous achievement, you may ask, but this is the very point that I will come to: The AI currently open to users has been trained with only fifteen thousand images, while the more enhanced future version to be available soon has been trained with a thousand times (fifteen million) visuals. Even the current AI performs OCR amazingly well, and the results appear quite fast. One of the most important things is that FYE's AI will be independent instead of relying on GPT. It also has the feature to instruct the AI to improve the descriptions or directly send the photos to volunteer humans to request a more accurate description from them should you need one. You can also access other stuff like ranks and leaderboards, and articles on how to use the app, which might not have been translated into English as of yet. The video description feature in particular will revolutionize how you watch movies and videos. It's something I had suggested even before the app was born, about one and a half year ago, and one of the many things that I have suggested up until now, so I am particularly excited to witness it become a reality.

Yes, the FYE app has appeared on the appstore and it is at version 1.0.2 (of 9 November), while the original From Your Eyes app is still on the appstore with version 1.0.21 (from mid-September).
For clarification:
- the very exciting new features you write about aren't available yet in the FYE app, are they? They will rather be gradually introduced later on?
- right now both apps work and do the same thing? The description of the FYE app is at present identical to the original From Your Eyes app.

I signed up, took some pictures and it said, unable to get picture in your AI language, or something like that.

It was a picture of a bottle of mouthwash. Then I tapped on make this description better with volunteers but there was no indication they were working on it.

Also the video feature isnt to do with youtube or stuff like that, at least at the moment it isn't, you can take a 1 minute video of something and it will try and describe it.

If it can describe video that would be cool but it would have to describe it real time which I don't think is possible just yet.

I've deleted my account and if I hear from others I'll try out the ap agian but I'm not impressed just yet.

OpenAI is reported to have been the target of a DDoS attack by a group of so-called hacktivists called Anonymous Sudan after the DevDay event where Sam Altman announced that they had surpassed 100 million users, and new features like GPT4 Turbo. From Your Eyes does have its own AI trained with 15 million images, but it still relies on OpenAI's language model to provide the descriptions. I've personally learned from the CEO that they're looking for alternatives that should face outages less often.

I'll personally leave this app alone. If others like it, great! But it's not for me.

Yeah, still not impressed, and really don’t think I’ll ever be, the way it sounds… Still haven’t been convinced to re-download it after all the problems I had before.

And the problems I mentioned are with GPT itself, so you can't blame that on FYE.

But at the moment it doesn't have things i'd want/need.

I installed the FYE app. Logging in was a bit of a challenge, as I thought that I could just log in with the account I had created for the From Your Eyes web app back in May. But this proved wrong, as it turned out that I had to set up a new adcount for the FYE app. This was quite hard to find out as the FYE app just read a Turkish text whenever I tried to log in with my earlier account and furthermore this text wasn't displayed, only said by VoiceOver. So I had to somehow catch it and spell out just by ear to have it machine translated on my laptop.
Based on my first experience with the FYE app, I have these questions:
- are the descriptions only available in Turkish for now? The language option in my profile is set to English and the user interface is (mostly) in English, with some untranslated bits here and there. But I get the description in Turkish.
- I presume that the volunteer descriptors aren't available 24/7
. Is it so? I think that is why the descriptions I have sent for improving hasn't been touched yet.
- in what language I can guide the AI for now?
- I presume the new AI trained on 15 million visuals isn't in place yet. Is it so? I ask this because I took a very simple image of my TV screen and both Be My AI and Seeing AI described it as a white logo (a silhouette of a bird) and two simple words (which they also read and also told their position) on a red background. But FYE was talking about an outdoor setting, trees, a bench and someone in a blue vibrant outfit. As if it was looking at a completely different image.