June 21, 2025
so i fixed the errors with the ai thingy and noww itss workingg :)
the issues was navy api's sec blocks any hosting's ip so i added diff diff headers and other stuffs to bypass it :>
alr.. ive uplaoded demo videos removed useless files that i got from examples and yea it was tested by my frnd so i think its pretty much working and whats now? ive uploaded all the codes to github ! :))
guess what? after working for hours its donee !! :))
that dma overflow issue was solved via removing live streaming as said earlier, now the issue was aftr clicking img and then uploading it the response wasnt tht accurate so? so i researched a bit and found that instead of uploading base64 encoded img i should upload it to thier server first which will give an upload id then i can use that upload id to label the img and it workeddd :))
oh well.. everything was wrking as planned after way too much fixesss but the dma overflow was still persistent , and i decided to remove the live streaming things so it wont load the ram and guess what it worked? but- NOW THE API RESPOSE IS ALL THE TIME SAME NO MATTER WHT THE IMG IS i think its cus of base64 encoding isnt working ryt? maybe its broken i need to find a way to fix it ;-;
uh.. my one esp32 got dead ;-; and THE NEW ONE IS SPAMMING DMA OVERFLOW LIKE WHT LIL MAN AHHHH ;-; i just addwd an opt for live streaming the cam seems like it wont work ;-;
this project - AHH its deff not my kind but yea i thought i should make some iot projects thats why i took this one..
today i bought items i.e, esp32 cam module.. pir sensor.. reed swtich...
and all this time i was just trying to test them all if they working or not make trying diff diff things the problems i faced?? just dont ask ;-;
the camera IT WAS SPAMMING DMA OVERFLOW in my terminal if thats not enough it was like cam init faild and stuffss so it wasnt so much of making till time.. its full of eorrrors lol will try to fix it asap.
It’s a smart IOT-based security system built on the ESP32-CAM board. Whenever motion is detected via a PIR sensor or the door is opened via a reed switch, the board captures an image. That image is then sent to an AI (Imagga API) to detect whether it’s a human or an animal. Based on that, it sends alerts on Telegram complete with the image and what it detected. No bullshit, fully wireless, real-time alerts. Built to be fast, low-power, and actually usable.
fixed the ui for confession listings no more bugss!!jhi
AHHHH finally… it’s ALL done!
Just fixed a bunch of bugs and polished everything. The biggest drop? VIBE & MATCHING
Now Xander lets you:
Fill out a vibe profile with aesthetic questions
Get matched with others who share similar energies
Use /reveal to connect only if both opt in
Admins can now force matches, remove profiles, and view stats
I also refactored the whole codebase no more giant index.js mess! 😮💨
Every feature is now modular, organized, and scalable.
Xander is now your go-to AI-powered daily Discord buddy: translator, note-taker, reminder setter, vibe matcher, you name it.
Time to vibe, connect, and conquer Discord. Hope y’all enjoy! 🖤
added some admin lvl cmdsss :)) to our botss
i'd love if u guys gimme some suggestions for more feature :3
FROM PAST 2HRS IM TRYING TO BUILD A THING VIBEMATCHER more like a frnd finder but these sh*tty errorss ahh
1: Built the entire Xander Discord bot from scratch today
2: Added all core features, AI commands, notes, reminders, and a fancy help system
3: Cleaned up the repo, removed useless folders, and wrote a pro README
So its pretty much ready to launch but i'll look into more how do i expand it <3
Xander is a multifunctional Discord bot powered by NavyAI GPT-4o, offering translation, AI chat, notes, reminders, vibe-based friend matching, and more all accessible via slash commands in DMs and servers.
ctfhost
is a clean Bash utility I built to solve a simple but annoying problem: manually editing /etc/hosts
for every TryHackMe or HackTheBox machine. It lets you instantly add, remove, list, or flush .thm
/.htb
DNS entries with style—emoji banners, colored output, and even your current VPN IP shown with each entry. I packaged it into a .deb
file for easy installation and made sure it feels like a native Linux tool. Built with love (and frustration), it’s now open-source and ready to be part of every CTFer’s toolkit.
I added About Pageeee
Guess What ?? I HOSTED IT AND THE APP IS FREGGIN LIVEEEE :D
todayy all i did is ui ahh why its not my expertise ;-;
btw.. today onli i also added firebase real time databasee
Started Building things in react.. till now discord auth is done and like basic auth workflow is working.
A modern, anonymous confession platform with a beautiful glassmorphism UI, animated backgrounds, and real-time social features. Built with React, Tailwind CSS, and Google Sheets (via SheetDB.io) as the backend.
DevLog #1 – Launching Byakugan
Today, I officially began building Byakugan, my dark-themed, AI-assisted recon tool made for bug bounty hunters and cybersecurity enthusiasts like myself.
The core idea is to automate recon workflows — subdomain enumeration, probing for live hosts, capturing screenshots, and using AI to instantly analyze risk, suggest potential exploits, and even tag domains for bounty potential. Imagine having a hacker’s sixth sense — that’s what I’m aiming for.
Here’s what I accomplished today:
• Built the CLI core in Python
• Integrated subfinder for subdomain discovery
• Used httpx to detect live hosts
• Wired in gowitness for screenshot automation
• Designed a clean output structure for storing results
• Connected the Gemini API for AI-generated recon summaries
I faced some annoying bugs like Kali’s pip install restrictions, gowitness misusage, and Gemini API rate limits, but I worked around them all — including setting up a proper virtual environment and caching outputs to continue from previous scans.
This is just Phase 1. Up next is adding more tools (like Amass, crt.sh), improving scan persistence, introducing multi-AI fallback (Grok > GPT > Gemini), and working on an optional FastAPI + Tailwind dashboard.
Byakugan is more than a tool — it's going to be the hacker’s all-seeing eye. Stay tuned.
— Lav (aka kairav | KafeStack)
Byakugan is a dark-themed, AI-augmented recon framework for ethical hackers and bug bounty hunters. Inspired by the all-seeing dojutsu, it automates subdomain discovery, live host detection, visual intelligence, and threat analysis using real-world tools (subfinder, httpx, gowitness) combined with AI models like Gemini and GPT. With a CLI-first workflow and optional FastAPI + Tailwind dashboard, Byakugan gives you a hacker’s all-seeing eye — spotlighting high-risk targets, suggesting exploits, and tagging bounty-worthy domains instantly.
This was widely regarded as a great move by everyone.