MirrorArc

INDEX (landing page)
Relationship Map (knowledge network visualization)
File metadata

Inspiration

AI generated too many documentations and misc. files for me. And I also referenced a lot of resource from online, some are in .docx, .xlsx, .pdf etc. I got so lost. I had to create a tool for myself to learn about what AI created, to understand all the docs and manage them.

What it does

It resolve the AI slopy and explosion of documentations, help user find the easy path to organize all the docs, read and interpret them in logical sequence and manage them in a portal

How we built it

mostly vibe coded with CodeX, tested with own dogfood documentaiton.

Challenges we ran into

the docs are different from data, and also very different from code. the procedure of handling different types of docs are challenging, luckily Markitdown helped a lot. and the LLMWiki is inspiring for the management part.

Accomplishments that we're proud of

The idea of Mirror is gold, inspired by SQL DB log shipping (replica db) The idea of growing Knowledge base is cool (inspred by LLMWiki) we combined both and created the multi-layer mirror, it is awesome.