mostafa kh

validating before building is the smart move. too many founders skip this and waste months on ideas nobody wants. community feedback is useful but curious how you keep the quality high. is there any filter on who gives feedback or is it open to everyone?

tried the scan on my site and got a score of 45. but thats all i got, just a number. no breakdown of what's wrong, no hints on how to improve, nothing to act on. felt like a dead end. the potential is there but if you want people to upgrade, you need to give them a taste of the value first. show me one or two specific issues so i understand why i need the full report. a score alone doesnt push...

this is the kind of simple tool that should exist but nobody built until now. dont need teamviewer or anydesk just to pause any running app from another room. rust rewrite under 5mb is impressive. most electron apps for simpler stuff are 200mb. the react sdk is a nice touch. gonna try building a simple dashboard on a spare pi.

the idea of splitting read and write between models is interesting but im not sure it always works. sometimes understanding the problem needs the smart model too. a cheap model might miss context that a bigger one would catch. does it handle complex multi file bugs well or is this more for simpler tasks?

this would have saved me so much time. spent days working with claude opus, chatgpt 5.2 and claude cli trying to figure out why ai bots couldnt read my blog pages. just scanned my site and got clear answers across all 7 dimensions in seconds. no more guessing. the llms.txt check is great. most people dont even know this standard exists yet. free and simple, exactly how tools should be. wish i...

2 minutes for a studio quality render is impressive if the output actually holds up.

the markdown native approach is smart. api docs and tests living in the same file that git tracks properly is something postman never figured out. the block composition idea is interesting too. treating headers, auth, body as reusable pieces instead of one big request object makes more sense. does it support importing postman collections?

tried to check it out but landed straight on a login page. google signup didnt work either. you say its free forever get started but users cant even see the tool before signing up. letting people explore first would help a lot with first impressions.

this goes way deeper than a typical vehicle tracker. the predictive modeling for resale timing is interesting, most people guess when to sell instead of using data. the fleet and auto lender angle is smart too. b2b is where the real revenue is. curious about the data sources for market signals. are you pulling from auction results, dealer listings, or something else?

the idea of finding leads while you sleep is appealing. manual searching on reddit and twitter is a time sink. curious about the ai intent analysis. how does it decide if someone is actually looking for a service vs just complaining or asking a general question?

the spatial thinking point is real. my brain works better when i can see things laid out visually instead of buried in folders. how does this compare to tools like miro or figjam? whats the main thing that makes lokryn different?

honestly surprised this is so low on the list. this solves a real gap that nobody else is addressing. google built a2a for agents to talk. anthropic pushed mcp for tools. but both assume you already know where the agents are. this is the missing discovery layer. using mdns is clever too, same tech that makes airprint and spotify connect just work. hope more people notice this one.

the table and inline css pain is real. spent way too many hours debugging emails that looked broken on outlook. does the html output work well with transactional email services like resend or sendgrid? or is this mainly for marketing emails you send manually?

gui for training models without cli is a nice idea. lots of people get stuck on environment setup before they even start. curious about the cloud gpu part. which providers does it support? and how does billing work?

this is needed. right now finding good skills is like digging through github repos hoping something works. how will you rate them? user reviews, testing yourself, or some kind of scoring system? the hard part is keeping it updated as skills change fast.

the bookmark trick is clever. most url tools make you copy paste back and forth but this keeps you in flow. do you support batch editing? like changing one param across multiple urls at once?

having backend, auth and deployment all in one place is the real value here. jumping between lovable and supabase and vercel gets messy fast even for simple projects. can you eject the code if you want to move off the platform later or is it locked in?

looks clean. curious how you position this against just asking chatgpt to translate into multiple languages at once? whats the main reason someone would switch?

the visual automation and spam checker before sending are nice touches. most dev focused email tools skip these and expect you to use something else for drip campaigns. how does the notion style builder work? can you mix markdown and drag drop blocks together?

i was browsing copart listings the other day, testing my own extension. found a car that looked clean, photos seemed fine. ran the damage analysis anyway just to test the prompts. ai caught something i completely missed . damage that wasn't even listed. and i've been buying and flipping cars for years. that's when i realized this tool actually works. it sees what we miss. if you're a serious...