AI testing storyteller. Writing the 36 Stratagems series — ancient Chinese military tactics meet modern AI system meltdowns. 16yr QA → building AI test framework.
On-device editing is compelling not only for privacy, but for a simpler trust story: the user can understand where the pixels went. I’d make that visible in the product with an offline indicator, a clear export path, and a documented list of features that ever call a network service. Privacy claims are strongest when they are inspectable behaviors.
I research AI tools, APIs & agents — comparing specs, pricing and real user reports — then publish honest rankings so you don't waste money. Research-first, never sponsored fluff.
The scoreboard is the part that makes this believable — you hand Canva the template round and Photopea the depth round instead of claiming APIC wins everything. The 'fit this file size' compressor is the bit I'd actually steal, binary-searching quality beats guessing a slider every time. One thing though: export needing a Google sign-in is the one line that undercuts the privacy story. Is that a forever thing or an anti-abuse stopgap?
like I've added export needed google sign in( firebase auth) just to count my users, and for sending them emails for welcome, updates that's it. also it never uploads your uploaded image to any server that's what I'm pointing or denoting as privacy first, otherwise there is just one cons that I've said publicly that you need to sign up for export... sorry bro for that cons 😔
AI testing storyteller. Writing the 36 Stratagems series — ancient Chinese military tactics meet modern AI system meltdowns. 16yr QA → building AI test framework.
Haha, looks like my mouth is blessed! 🤣I remember saying your articles would keep hitting the hot list — and it actually came true! Congrats! It’s all because you keep putting out such high‑quality work.👍️
Developer, maker, and perpetual project starter. I build web apps and dev tools mostly, things that I think are really useful and should exist and be simple to use. I also like to experiment with AI.
This is seriously impressive. I actually built a browser-based photo editor myself that uses AI models, so I know how much work can go into something like this. Mine has been sitting a little neglected lately, but it's a really fun kind of project to work on.
I especially love that you've put so much thought into privacy and keeping things on-device. I also appreciate that you're open about the limitations and what could be improved. That's such an important part of building software.
It's really cool to see someone so young already creating and sharing projects like this. Keep experimenting and building things. Awesome job on this.
The on-device-first architecture is exactly the story we hear repeatedly from desktop users who won't upload photos to a cloud remover. ONNX Runtime Web + WebGPU is a nice fit here, and the same model graph ports cleanly into an Electron app because the renderer gets the same WebGPU/WebAssembly stack — the only thing you add is Node-side file I/O, so you can batch over a folder of RAWs without a drag-and-drop step. The one caveat from the desktop side: a multi-hundred-MB model bundle you download lazily in the browser becomes an install-size decision when it ships inside a signed DMG/EXE, so the three quality tiers (small / medium / full) you already expose should probably be pre-built as optional download packs rather than bundled in.
We're a place where coders share, stay up-to-date and grow their careers.
@technogamerz didi I've upgraded APIC WEB more so can you please try it?
Are bhai aaj diwali hai thoda rukiye na.Happy Diwali ❤️
yeah, happy diwali.
Happy Diwali me aapko ek funny baat batana chahati hu batao kya?
Yes why not ?
Me jab bhi Bahar jati hu toh log dukandaar se kahte hai bhai scanner do kya voh log sahi hai🤔
HAPPY DIWALI @technogamerz @effessdev @xulingfeng @koda2026 @merbayerp @star123
HAPPY DIWALI ❤🔥❤🔥❤🔥
Happy Diwali ❤️
Happy Diwali ！
So...what is Diwali？？？😂
Go search on UC browser lol 😆 😂
Bhai wow kitne patakhe jala rahe hai log 😮 me bhi jati hu bye 👋🏻👋🏻
On-device editing is compelling not only for privacy, but for a simpler trust story: the user can understand where the pixels went. I’d make that visible in the product with an offline indicator, a clear export path, and a documented list of features that ever call a network service. Privacy claims are strongest when they are inspectable behaviors.
Point to be noted thanks bro for the comment
The scoreboard is the part that makes this believable — you hand Canva the template round and Photopea the depth round instead of claiming APIC wins everything. The 'fit this file size' compressor is the bit I'd actually steal, binary-searching quality beats guessing a slider every time. One thing though: export needing a Google sign-in is the one line that undercuts the privacy story. Is that a forever thing or an anti-abuse stopgap?
like I've added export needed google sign in( firebase auth) just to count my users, and for sending them emails for welcome, updates that's it. also it never uploads your uploaded image to any server that's what I'm pointing or denoting as privacy first, otherwise there is just one cons that I've said publicly that you need to sign up for export... sorry bro for that cons 😔
Haha, looks like my mouth is blessed! 🤣I remember saying your articles would keep hitting the hot list — and it actually came true! Congrats! It’s all because you keep putting out such high‑quality work.👍️
Thanks bro still I'm trying to match quality of your stratagem series
This is seriously impressive. I actually built a browser-based photo editor myself that uses AI models, so I know how much work can go into something like this. Mine has been sitting a little neglected lately, but it's a really fun kind of project to work on.
I especially love that you've put so much thought into privacy and keeping things on-device. I also appreciate that you're open about the limitations and what could be improved. That's such an important part of building software.
It's really cool to see someone so young already creating and sharing projects like this. Keep experimenting and building things. Awesome job on this.
The on-device-first architecture is exactly the story we hear repeatedly from desktop users who won't upload photos to a cloud remover. ONNX Runtime Web + WebGPU is a nice fit here, and the same model graph ports cleanly into an Electron app because the renderer gets the same WebGPU/WebAssembly stack — the only thing you add is Node-side file I/O, so you can batch over a folder of RAWs without a drag-and-drop step. The one caveat from the desktop side: a multi-hundred-MB model bundle you download lazily in the browser becomes an install-size decision when it ships inside a signed DMG/EXE, so the three quality tiers (small / medium / full) you already expose should probably be pre-built as optional download packs rather than bundled in.