leocerv

Im really excited about this new project of mine, im gonna double dip this with endpoint but ive had this idea for a while, basically its a legal doc generator based on the needs you have on your website/project/whatever.

I wanna focus on shipping the endpoints fast, but i do wanna expand it more in the future. for now ive set up the framework which is ruby on rails (i dont have much experience with it). so ive been struggling a bit, after watching some vids i think im getting the hang of it.

i also made a little logo for the project which im gonna use as the cover too, might remake it cus i didnt love it.

i made a new release, improved the readme a little bit more

so i found out a game changer, which is that in irohs gossip protocol, you can adjust the message size limit in the client, and with this i was able to get the best quality possible for the video, and since i adjusted the limit, the video was sent completely without errors.

so for this new devlog i tried to make a working version of a p2p videochat, but i just couldnt make it work, i tried everything until i downscaled the res of the video and thats when i was able to get it working, but it took so much time of beugging, reading, and trying new things only to fail.

i also did a little refactor to the chat.rs file but its something minmal, i spent like 30 mins doing that, most of time spent went down onto the videochat. i sttach a demo video of the app working

i started this project few hours ago but i wasnt sure if i was gonna ship it to som or not, so here i am posting a 8hr devlog, what i did so far was create a p2p chat using iroh and following their code while making adjustments.

I also made a module to retrieve frames from your camera and then process them and display that frame using halfblocks in your terminal, the result is attatched, its stable at around 60-70 fps which is a good number and the quality isnt that bad.

So the model and the whole app is now hosted on hugging face spaces, someone on the slack suggested this and it was a great fit for my app, everything is working fine now, including the video inference.

made a prettier, better and more complete readme for this project

Hey guys! so i just want to announce the first version of unsent canvas is shipped and ready for everyone to try it out, i put this on the slack first, like a week ago and i wanted to see how it performed, and i think it had positive feedback, so i finally built up the courage to ship this in som.

pd. i got rejected from hackmate lmk if this was deserved

hey guys! so i just wanted to share that ive been working a lot on hackmit puzzles, way more than shown right now on the devlogs, but the point is im doing great, as for now i have 4/6 CTFs done, once the deadline is over i plan to do some videos or guides on how i cracked them! im also #39 as for rn

TL:DR fixed the demo link deployment, you can see on the image its up on the correct link

fixed the nest problems i think, did a basic review on what some issues might be and apparently the site is running correctly, im not sure if it was me, or nest was down, but the point is the demo link works!!!

I added much more colors, also made an attempt at lazy loading and better note rendering performance which didnt go so well, but it helped me identify that the cursors were lagging the app, and that they were taking up so much of the message quota for the realtime db

TLDR;this update brings a safer, improved canvas.

The note hover animation is now more subtle to avoid interference with adjacent notes, and the note preview is fixed and matches the canvas, just much bigger and easier to read.

Content moderation is stronger with automated filtering and a report button in the preview. Each note tracks unique reports. Spam prevention is smarter, with better rate limiting, pattern detection, and IP blocking. I also refactored all API routes for a cleaner backend.

spent this time fixing an error, and i was fighting with imaginary bugs just because nest was failing. the i struggled with the auth and bot token handling, but the slack bot can now be downloaded in any workspace!!

Added some really important features!! including:
note cooldown (cant post a note for x time)
cursor tracker (tracks your cursor and shows it in the frontend of the other users)
collaboration on the canvas, the other users cursor and the note creation upload in real time on everyones frontend.

this app is looking really good so far, and im really excited to see what i can add or improve!!

Added the canvas!!!! in this canvas you can create the notes you want, everything is completely anonymous, but its not collaborative yet, although it is connected to the db

the db connection and the note creation is another thing finished!! imo everything regarding notes is almost finished, besides the creating cooldown i plan

worked on making a prettier and more informative readme

i focused on fixing some minor issues and making the index prettier

hey!! so ive just started with this project im excited about, one of the websites i really love is the unsent project i think it was a really cool idea and ive always wanted to do a clone, so thats what im gonna do, the only difference will be all the notes are going to be in a huge collaborative canvas, like pixel io (i also got the idea from there).

finally deployed on vercel!! since nest is having some serious issues rn

implemented some really cool ascii art using AI!!

worked on developing the web interface for the interpreter, and got it nice and runing finally!!! im gonna be working on making some ascii art cus i think its going to be cool.

THE GROOT LANGUAGE IS FINISHED!!
i added a lot of stuff, you could say i completely refactored the entire thing, commands have changed, you now only have 2 variables and one optional function, no more, but u get error handling and im working on making some test programs but its super hard to code in this language.

also the program passed all the parsing tests which mean its identifying each token correctly. i also added comments using #

This first 4 hours were spent creating the base and defining the different tokens for the groot esolang, for now i plan to use different punctuations to indicate loops, and create variables and other stuff, but i still have to finish writing the interpreter

got a working version of the model and the web interface deployed on nest, check it out at https://flowingvision.leonardocerv.hackclub.app/

the model inference isnt the best bc open vino is optimized for intel and nest runs on amd, nevertheless i think i did a good job and ill be ready to ship too!!

made the web interface prettier, and focused on making the model actually run the inference live on camera, it was such a struggle but now it works, next step is making the user able to upload a picture

i updated the readme and prepared all the links for the ship release!!

added a home page! from which you can install in different platforms, i included redirect urls for me to be able to share the slack version of the bot, eveyrhting hosted on nest!

hey guys so i finally deployed my bots using nest!!, and i ensured there is consistency between all the bots, because each bot had like different naming conventions for the same command, or different responses to the same tool, so yeah thats what i did. i also tried improving the regex patterns because i think they could be better since im a begineer on regex and lambda but it didnt work so i just kept the latest working version

there is now support for the bot on telegram!! i think im almost ready to ship this, i still feel like making some improvements and definitely have no idea on how to deploy this.

This bot now has support for slack!!! while its not exactly the same functionality, and is still not very polished, i think this is a pretty good base for this bot, this was much more easier to make than the github version, idk if it was because im new and the first one is harder, or because its easier to work with bots in slack.

you can now react to a message and the bot will send a dm with t
he translated timezone!!

the bot can now accurately convert messages using ephemeral message to the users set timezone or to a specified timezone, i wanna work on a react on message command next.

Hey guys so im excited for this new bot, as for now im starting with the discord platform and i think its going well, i still havent donde much progress but i got a functional bot, just not with the actual functionalities i want

Hey guys so ive been trying to make a web interface for the ai model, but ive ran into some issues, im still gonna share the progress of what i have as for right now, the ultimate goal is for the model to be able to run on your camera without you needing to install anything.

i implement a collection creation functionality, which can also link prompts to this new object created, the prompt view for each collection doesnt work tho so im gonna be working on fixing that

The dialog for the prompt creation is so much cleaner and pretty now, i think ive made good progress with this, the search functionality isnt working yet but i plan on improving that next, along with implementing the collections.

finished CRUD for my prompts, i also added new data models for my supabase connection and added tables here too. Fixed an issue with the prompt edit. next steps is fixing real time update of the prompts, since rn you have to reload the page to see changes

I started with the database integration into the app, but it didnt work as expected, i took way too much time trying to fix bugs, so im gonna focus on organizing my ideas with diagrams and project management tools to have less problems when trying to code everything, specially since theres a bunch of data, tables and relations that i cant keep track of

In this new update i implemented major improvements to the prompt editing workflow and marketplace navigation. Added new components for editing and displaying prompts, introduced a dedicated marketplace layout, and enhanced UI consistency across pages. Refactored several files to improve code organization and maintainability.

This update introduces a comprehensive UI for prompt management, including new components for listing, previewing, and organizing prompts. The /prompt route now has its own layout, and the prompt page has been refactored for better structure and maintainability. Several reusable UI components were improved, and a custom hook was added to streamline prompt data handling. These changes lay the groundwork for a more robust and user-friendly prompt workflow.

Reorganized the app's routes and components for better modularity and scalability. Added dynamic pages for marketplace, prompt, and user details. All feature pages now live in their own folders for easier maintenance.

I migrated my auth system to supabases auth, and its super convenient. i ran into some issues, but i hope that in the long run is gona be easier to mantain and also to debug, next steps is allowing the user to interact more with the app, in the sense of storing prompts, creating collections, etc.

I created the nextjs app and setup the tools and themes according to my idea, if its a wisp, the app needs to be wisp colored right?

I also added a simple login functionality (which took me more than it shoulve), including the signup and logout but its not complete yet, theres some bugs in the login and no database save which ill solve later.

DETECTION MODEL RUNNING! 💧

TL;DR: My AI leak detection model (made from scratch) is ready for anyone to check out, you can test it cloning my repo for this project, ill focus on adding a readme so the setup is a breeze for anyone. Feel free to hmu regarding issues or questions.

Hey people, this is my first of many devlogs for this project, i just finished a MASSIVE milestone in my Flowing Vision project and I'm REALLY EXCITED about the progress!

Improvements summary:
AI MODEL:
✅ Successfully implemented the leak detection system using OpenVINO for a faster inference
✅ Created a smooth pipeline from PyTorch -> ONNX -> OpenVINO using python programs
✅ The model now reliably detects water leaks with confidence tracking over multiple frames
✅ Added a wack confirmation logic that requires 20+ consecutive frame detections for a leak to be valid

DATABASE:
✅ Connected to MongoDB for persistent leak storage
✅ The model can still run without the need for a mongo connection
✅ Automatic database recording of confirmed leaks
✅ Base64 encoding for image (ss) storage
✅ Each leak gets a unique timestamp-based ID

VISUALIZATION:
✅ Implemented live camera feed processing (its low fps sometimes)
✅ Added real-time bounding box visualization
✅ Created on-screen status indicators for DB connection
✅ Added frame counters for continuous leak detection

WHAT'S NEXT?
I'll now focus on creating a better README with detailed setup instructions for the model testing. This will make it easier for anyone (including me) to get the system up and running quickly.

After that, I'm finally diving into building the Django app to create a proper web interface for this system. Time to move on to a web app that can display leak data, historical records, and the leak screenshots!

💧💧 Any feedback is very much appreciated! 💧💧