What If Your Chat History Could Tell You the Truth?

#devchallenge #muxchallenge #showandtell #video

DEV's Worldwide Show and Tell Challenge Submission 🎥

This is a submission for the DEV's Worldwide Show and Tell Challenge Presented by Mux

What I Built

I built ChemistryCheck, an AI-powered tool that analyses chat exports from WhatsApp, Telegram, and Instagram to show what's actually happening in a conversation.

You upload a chat file, and it breaks down communication patterns: who initiates more, response delays, effort imbalance, message volume, word usage, emojis, and consistency over time.

The goal is simple: turn vague feelings into clear signals.

My Pitch Video

In the video, I explain the problem, why people miss red flags in chats, and how ChemistryCheck makes those patterns visible instead of emotional.

Demo

Live app: https://chemistrycheck.vercel.app
GitHub repo: https://github.com/ArjunCodess/chemistrycheck

Here's a sample chat analysis link: https://chemistrycheck.vercel.app/analysis/00d311b9-8619-4bae-8367-029aea4eeae6

How to try it yourself:

Sign up or log in
Upload a WhatsApp (.txt), Telegram (.json), or Instagram (.json) chat export

[for the judges, I have uploaded three of my chat exports here so they can download it and test the product right away]

Wait for the analysis to complete
View insights on the dashboard

The Story Behind It

This project came from personal experience.

When a conversation starts hurting, people often rename the pain. We call it "trauma bonding," "mixed signals," or "bad timing." But the red flags usually show up much earlier in the chat itself.

Uneven effort. Long gaps. One person is always initiating. Warm one day, cold the next.

I built ChemistryCheck because I kept rereading chats and still couldn't tell what was real. I wanted something that looked at patterns instead of excuses.