DEV Community

Purvi Jain
Purvi Jain

Posted on

Title: Meeting GPT™ — I Built an AI That Turns One Meeting Into Six Published: False

April Fools Challenge Submission ☕️🤡

Every team has that one meeting.
The one that could have been an email.
But instead… it becomes:

A follow-up meeting
A sync meeting
A pre-meeting before the next meeting
And somehow… still no decision

So I thought:
What if we stop pretending meetings are efficient?
What if we optimize the chaos instead?

Ladies and gentlemen: Meeting GPT™.

“Let’s circle back… forever.”
**
What I Built**
Meeting GPT™ is a web app that takes a simple meeting topic — and expands it into a fully over-engineered, never-ending corporate meeting ecosystem.

You enter something like:

“Discuss project deadline”
“Fix login bug”
“Plan college fest”

And the system generates:

A chain of unnecessary meetings
Corporate jargon-filled agendas
Action items that go nowhere
And zero actual decisions

Demo
Try MeetingGPT™ Live →

Here’s what happens when you use it:

Enter your meeting topic
Type something simple like:
“Fix login bug”, “Discuss project deadline”, or “Plan event”
— because that’s all it takes to start the chaos.

The system begins scheduling
A loading sequence kicks in with highly professional (and deeply concerning) messages:

“Identifying key stakeholders…”
“Adding unnecessary participants…”
“Aligning on alignment…”
“Scheduling pre-meeting for main meeting…”

The meetings multiply
Your single topic expands into a full calendar crisis:

Initial discussion call
Pre-discussion sync
Cross-functional alignment
Follow-up meeting
Post-meeting debrief
Final review (that decides nothing)

The jargon escalates
Every meeting includes phrases like:

“Let’s take this offline”
“We need more visibility”
“Circle back later”
“Leverage synergy moving forward”

Time disappears
The Time Waste Calculator updates:

“Total time spent: 30 minutes”
“Actual progress: 0%”

Decision avoidance achieved
You receive a list of action items:

“Discuss in next meeting”
“Revisit later”
“Create doc for future discussion”

Code
(PurviJain027/Dev_challenge_Aprilfool)

How I Built It

  • Stack: HTML + CSS + Python (via PyScript / Antigravity) + Gemini API (Google AI)
  • No backend. No server. No actual productivity. Runs entirely in the browser.
  • The interesting technical choice here is PyScript — Google's Antigravity project that lets you run Python directly in the browser via WebAssembly. So the Gemini API call is written in Python, not JavaScript. This means:
  • No Node.js, no npm, no unneeded dependencies (unlike our meetings)
  • Python's google-generativeai style calls right in the browser
  • The user's API key never leaves their machine — it goes directly from browser to Google's API

Prize Category
Best Google AI Usage — Gemini powers the entire corporate-jargon generation engine, and the app uses PyScript (Antigravity) to call it directly from Python running in the browser — absolutely no backend required. I used Google AI Studio extensively to iterate on the system prompt, testing dozens of simple work tasks until the output hit that perfect, soul-crushing balance of unhinged bureaucracy and mock-authoritative scheduling. The whole stack — from browser Python straight to Gemini's language model — is Google AI products all the way down.
**
What I Learned**
Building something intentionally inefficient teaches you a lot about what makes things feel productive.

Turns out: it's mostly presentation. The loading messages that say "Aligning on alignment..." make a simple UI delay feel like a cross-functional effort. The "Time Waste Calculator" — completely arbitrary — makes the lack of output feel quantified and measured. The mock-professional tone of the AI response makes total workflow paralysis feel like "strategic corporate alignment."

The real joke isn't the useless calendar blocks. It's that this is exactly how a lot of modern corporate infrastructure operates. Just with better PowerPoint slides.

That's a useful thing to know.
Trusted by 0 users worldwide. We are not responsible for any decisions made, overthought, or abandoned as a result of using this tool.

Top comments (0)