As someone who is blind and has always loved Japanese media, I've often wished there were a way to experience manga. Since manga is such a visual medium, it has traditionally been difficult for blind readers to access. This project is my attempt to make that experience more accessible.
This application is an accessible manga reader that uses AI vision models, including Gemini, Claude, ChatGPT to analyze and describe manga pages. The AI describes visuals and reads text following comic reading rules to help convey the story in a more complete way.
For transparency, I don't have a programming background. This project was developed with the assistance of Claude by Anthropic, which I used throughout the development process to build the application.
This is the first public beta release, so you may encounter bugs or unexpected behavior. Feedback and suggestions are always welcome, and I'll do my best to improve the application over time.
It's also worth noting that AI isn't perfect. Manga can be visually complex, and descriptions won't always be completely accurate. However, from my experience, the results are generally good enough to make following and enjoying manga series possible.
You can download the beta release here:
https://github.com/Shaima-Almarzooqi/Accessible-Manga-Reader/releases
I hope this project helps make manga more accessible and gives other blind readers another way to enjoy the stories they love.
By Shaima, 17 July, 2026
Forum
Windows
Comments
good work!
Thank you for the great idea of creating an accessible manga reader for blind and visually impaired people! I’ve always wanted to be able to read manga in an accessible way. How do I use the program?
Thank you
To use the program, download the .exe file from the link below:
https://github.com/Shaima-Almarzooqi/Accessible-Manga-Reader/releases/download/v0.11.1/AccessibleMangaReader-x64.exe
(That's the version for most Windows PCs. If you're on a Windows ARM machine, such as a Snapdragon laptop, download the arm64 file from the releases page instead.)
Open the file from your Downloads folder. It needs no installation and runs immediately. On first launch, Windows may show a SmartScreen warning that the app is unrecognised, since it isn't code-signed; choose "More info," then "Run anyway."
To read manga, the app needs an API key from one of the supported AI services. For most people that will be Google Gemini, which offers a free key. Other services work too if you have a key for them.
If you're not sure how to get a key, there's a full user manual built into the app: press F1 to open it, and it walks you through the process step by step. It also covers all the features and keyboard shortcuts.
If you need any help or have questions, feel free to email me at [email protected] and I'll be glad to assist.
A question
Where can we find manga that work with this application?
Where to Find Manga
I'm not sure if it's okay to direct users to manga websites here, but if you're not familiar with it, you can search for "MangaKatana" on Google. You can download manga from there in ZIP format which works in the app.
Other comics
Hi,
Firstly, congrats and thanks for such a great app. I'm primarily a Mac user but I'll try this app on my windows computer. My question is, is the app only supports manga or can we read other comics too?
Thanks in advance.
Thank you for this
I’ve been looking for a way to read manga for years! I’ve been using Seeing AI and be my AI to read comic books for the past year also, and that’s been working really well so I’m hoping this is similar.
Same Here!
That's actually what inspired me to build it. I had been using AI vision tools like ChatGPT to read manga, but constantly uploading a page or two at a time and writing prompts became tedious. I wanted something that could process an entire chapter or even a whole volume in one go, while also giving more control over how the content is read. Users still have the flexibility to give AI additional instructions before processing the chapter/volume to tailor to their preference.
The app works on the same general idea of using AI vision, but it's designed specifically for comics and manga. It also supports Western comics and webtoons—you can select the comic type in Settings → General, and the app will use the appropriate reading order. I haven't personally tested the experience with Western comics or webtoons yet, though, since my main focus has been manga, so if anyone tries it out and has any suggestion feel free to reach out.
About Gemini models
I use a free Gemini API key, and many manga or story chapters are very large, so I think the free API limit isn’t enough. The Gemini 3 model is the one that comes preconfigured in the app.
My question is this: if I switch to the Gemini 2.5 model, will the description quality drop significantly? If I use that model, I’d be able to describe more pages than with Gemini 3, but I’m worried about losing too much quality in the descriptions.
Interesting Tool
Even as a low vision user, I have found reading comic books and other similar type content rather difficult. So I've largely avoided reading it. I have started using AI apps like Be My Eyes and Ally to read comic books, but don't enjoy the rather tedius aspect of doing one page at a time, pasting in the same prompt each time, etc.
If this works with regular comic books too, I'm quite interested. I have a lot of TMNT comics I picked up a while ago that I'd love to read more easily.
About Gemini Models
Switching to 2.5 probably won't make much difference. Gemini's free usage limit is based on the number of requests you make each day, not how many pages or how much text is in each request. Since 2.5 Flash and 3 Flash have pretty similar daily limits, you'd mostly just be using a slightly less capable model without gaining much.
I think the issue is that you're using the default preview version (Gemini 3 Flash Preview), which usually has stricter limits than the stable release. I'll make Gemini 3.5 Flash the default in the next release so this shouldn't be an issue going forward. For now, try this instead:
1.
Go to Settings, click Refresh model list, and switch to Gemini 3.5 Flash (the stable version) instead of the preview one.
2.
Increase Pages per request to 8 or 12. Since the limit is based on requests rather than pages, you'll be able to process more pages each day without affecting the quality.
3.
If that's still not enough, you can add multiple API keys and switch between them. Each key has its own daily quota, so using more than one gives you additional requests once you've reached the limit on the first key.
App Support For Other Comics
Hi Jesse,
I'm glad to hear that! And yes, the app supports regular Western comics too, not just manga. It should work with TMNT and other comic styles as well, so hopefully it'll make reading them a lot more enjoyable.
re: About Gemini models
Thank you very much. I’ll try the tips you gave me. I’m also trying to create another free API key for another email account that I have.
re: About Gemini models
You're welcome! One more tip: you don't necessarily need another email account. On the same Google account, you can create a second project and generate a new API key inside it, each project has its own free quota, so that gives you more capacity too.
Vision Assistant Pro
How is that different from what I can do with Vision Assistant Pro and more?
How The App Is Different
That's a fair question. Tools like Vision Assistant Pro and Be My Eyes are genuinely great, I use and recommend them myself for all-purpose daily use and general image descriptions. The difference isn't that this app is "better," it's that it's designed specifically for reading comics.
General vision tools are meant to describe whatever image you give them, one image at a time. That's incredibly useful, but comics are a bit different. They have a specific reading order (right to left for manga, left to right for Western comics and webtoons), dialogue needs to be matched to the correct characters, sound effects and silent panels matter, and each type of comic has its own reading flow. To get consistent results from a general tool, you'd have to explain all of that in your prompt every time you analyze a new page.
With this app, you don't have to think about any of that. It already knows how each type of comic should be read, so you can just open the file and start reading.
The other big difference is how it handles larger comics. General vision tools work one page at a time, which is perfect for photos or documents. But a manga chapter might be 30 or 40 pages, and a full volume can be well over 200. This app processes the entire chapter or volume at once and saves the results, so you can read continuously instead of stopping to analyze each page.
So it's not really about replacing tools like Vision Assistant Pro or Be My Eyes. They're excellent for all-purpose daily use cases and general image descriptions. This app just focuses on one specific task, which is making it as easy as possible to read comics from beginning to end.