Inspiration

The inspiration was that while coding a lot of times we come across bugs and are not able to locate it or if someone uses ai to code as well the agent fails sometimes in identifying those small bugs so to help streamline the fixing process so that as a coder it is easier to locate or fix bugs we designed FaFi.

What it does

FaFi isn't your regular chat agent it is an agent that delivers directly you have a code with errors in it that you just cannot seem to locate FaFi does the trick. You upload your project logfile and codefile into FaFi through any of the 3 means we have provided: manual upload, local directory or github integration and it analyses your project in either the locate bug section or locate + fix section and it gives you the 3 guiding questions, the identified issue, the suggested fix and bug report. If you opted for locate + fix it also gives you the fixed section of the code along with the rest of the stuff listed above.

How we built it

We built the agent with the help of cursor for the code while the frontend was redesigned with antigravity and to generate the backend we used firebase. It consists of HTML, Java Script and Procfile languages.

Challenges we ran into

While changing the user interface the site became non functional all of a sudden due to the changes and we were not happy with the results so on realising the same we worker towards fixing it and luckily had the old code saved so we could go back to it.

Accomplishments that we're proud of

As the creator i feel our main accomplishment was that we integrated github into it though it is more of a creative thinking standard but it made me proud because it is a tool that comes handy when you dont have a pc with the code downloaded and you have the code on github saved you can easily use it to identiffy then and there instead of waiting to download and upload the actual files.

What we learned

Building FAFI taught us to turn AI into a structured problem-solving system, breaking issues into clear steps like analysis, fixes, and reports. We learned real-time backend handling with Firebase and the importance of clean UI/UX to clearly communicate complex AI outputs.

What's next for FaFi- Fast Fix Agent

We plan to integrate FAFI with platforms like GitHub to enable real-time code analysis directly from repositories and pull requests. Future improvements include smarter AI debugging, automated test case generation, and deeper integrations with developer tools to make FAFI a complete end-to-end AI debugging assistant.

Built With

Share this project:

Updates