Introducing Trott!
An organized landing screen for an organized app
Precise and succinct details of each video regardless of the genre
Search smoothly in natural language without friction
The universal problem you didn't realise you needed a solution to.
Our Adaptive and intelligent solution
The crazy 48 hour pivot!
The monetisation model
Our Unit economics
Trott's Future Scope
Thanks!

The Trott Story: From Digital Graveyard to Organized Paradise

🎬 Inspiration

Picture this: We're in Goa on a vacation, woke up late so rushed to get ready and are now sitting in the car and I ask "so, where's the amazing beach that you couldn't stop talking about?", after frantically scrolling through Instagram saves and our whatsapp chats for close to 30 minutes and a mini fight, we decided to order in and stay at the bnb!

That night, we had The Realization. Between us, we had a huge number of saved reels and shorts - travel spots, recipes, workouts, random products we'd "definitely buy someday." Our saves weren't memories or todos to fulfil; instead, they were a digital graveyard where good intentions went to die.

The final straw? My friend texted: "Did you guys visit that mud bath place in Goa? The one I shared on Instagram?" I knew EXACTLY which reel she meant. Could I find it? Absolutely not. That's when Sudipta looked at me and said, "We're developers. Why are we living like this?"

🚀 What It Does

Trott is your saved videos' personal assistant with a photographic memory.

Share any video from Instagram or YouTube to Trott, and our AI:

Figures out what kind of video it is (travel, recipe, workout, etc.)
Extracts the actually useful stuff:
- Recipes: Ingredients, steps, cooking time (no more pausing videos 47 times)
- Travel: Exact locations with Google Maps links & built circuit maps (goodbye, "somewhere in Bali")
- Workouts: Exercise lists, equipment needed (finally organize that fitness folder)
- Products: Prices, where to buy (your wishlist, but organized)
Makes everything searchable with natural language ("that pasta with the weird green sauce")
Lets you chat with an AI that remembers every video you've saved

Basically, we turned your chaotic saves into an organized, searchable second brain.

🛠️ How We Built It

August 15-16: The Telegram Bot Era

Started as a Python script + Telegram bot for travel videos
Added Gemini AI for location extraction
MongoDB for storage | Commits: 29

August 17-25: Backend Foundation

FastAPI + Firebase auth + Docker
Still just a travel app

August 26: Flutter Awakening

First mobile commit - goodbye Telegram!
Commits: ~50

August 27-September 5: Getting Fancy

Vector search, RAG chat, push notifications

September 6-10: RevenueCat Integration

Full IAP infrastructure
Still travel-only

September 15-16: YouTube Expansion

Added Vertex AI support
42 commits in 2 days

September 20-22: The 48-Hour Pivot

Users wanted recipes, fashion, and products
Hour 0-24: Sudipta rebuilt the entire AI pipeline for multi-genre
Hour 24-48: I redesigned all UI with genre-specific cards
First recipe saved = tears of joy
Stats: 27 commits, 150+ files, 0 sleep

September 23-28: Polish & Ship

Gemini 2.0 upgrade
47 bug fixes
Shipped v1.1.6

Total: 495 commits in 6 weeks | 2 architecture rewrites | 1 pivot that almost killed us

Tech Stack That Survived:

Backend

Framework: FastAPI + Python (async framework)
Cloud / Hosting: Google Cloud Platform (GCP)
- Cloud Run – serverless hosting & auto-scaling
- Cloud Storage – video storage
- Cloud Tasks – asynchronous background processing
- Artifact Registry - for Docker images
- Vertex AI
Database: MongoDB Atlas (free tier, 512MB)
Payments Management: RevenueCat

Frontend

Framework: Flutter 3.3+ with Dart
State Management: BLoC pattern
Dependency Injection: GetIt
Animations: Rive
Networking: Dio with caching
Firebase Suite: Auth, Messaging, Analytics, Crashlytics
Payments: RevenueCat Flutter SDK
Static Site / Marketing: Astro

Infrastructure / DevOps

Website Hosting: Vercel
Serverless Backend: Cloud Run for auto-scaling
CI/CD: GitHub Actions, Cloud Build
Email Automation: Brevo

😅 Challenges We Ran Into

Challenge 1: Saving Without an API (and Waiting Without Waiting)

Problem: Instagram doesn’t let you access saves, so we had to hack around it with the share sheet.
But then: Early version made users sit and wait for processing, unacceptable if you want to keep scrolling.
Solution: Built an async system with background processing, status polling, and upfront charging with adjustments later. A lot more work, but seamless for users.

Challenge 2: The Pivot Panic

Users wanted: Recipes, workouts, products, educational content… and YouTube videos too
We wanted: To keep it simple
Reality: We gave up and built everything

Challenge 3: Pricing Tug of War

Problem: If the extracted data wasn’t accurate, Trott would fail its entire purpose, but we also had to keep it affordable for people like us
Solution: Balancing accurate AI extraction + cost efficiency. Lots of optimization, late-night infra tweaks, and pricing headaches later… we made it work

Challenge 4: The AI Spoke Only English

Problem: Indian users were sharing Hindi, Tamil, Telugu content
Our 2 AM Solution: Multilingual support speedrun
Result: AI now speaks more languages than we do

Challenge 5: Cross-Platform Chaos

Problem: Trott runs on iOS, Android, and web. Testing & keeping it stable nearly broke us
Solution: Long nights of debugging, lots of “works on my machine” arguments, and finally… smooth sync across platforms

Challenge 6: The Marketing Bump

Problem: We're total noobs at marketing stuff like ads, campaigns, and reach all felt like alien territory
Solution: A lot of trial-and-error, learning on the go, and more mistakes than we’d like to admit. Still figuring it out, but getting better every day

🏆 Accomplishments That We're Proud Of

Did It All for the First Time: Flutter, iOS App Store, IAPs, building in public, even making videos with zero prior experience, all learned on the fly.
Shipped Anyway: Pulled this off while holding down 9–5 jobs. Sleep became optional; shipping didn’t.
The 48-Hour Pivot: Users wanted more, so we tore the app apart and rebuilt it in two days, and somehow it worked!
Pricing Without Pain: Went with pay-per-use instead of locking people into subscriptions. Subscription fatigue is real! and at just ₹99 for Starter, ₹299 for Standard, and ₹499 for Premium, it’s insanely affordable. The Premium plan is so generous that most users can probably cover the entire year for ₹499. Tons of testing, R&D, and iterative development went into making this possible.
600+ Videos Organized: Friends actually use it, and they keep asking for more credits.
Best Part: Hearing “This is exactly what I’ve been looking for!” from early users.

💡 What We Learned

Technical Lessons:

Your first idea is probably not it (travel app → universal organizer)
Things can always be improved and optimized
Users will use your app in ways you never imagined!
Building in public = being full time QA + constant feedback + building motivation

Life Lessons:

Two teammates + one problem + 60 days = chaos but also magic
Your personal frustrations might be universal problems
Sometimes the best features come from users breaking your app
When you love something, you let it grow
Coffee has a limit. We found it.

Business Lessons:

People hate subscriptions more than we thought, especially for something that is new
Learn Marketing and sales basics and important markers and metrics
Revenuecat experiments are better than guessing
If users are asking for something repeatedly, just build it and validate it
Trying to give true and genuine value to your user through your product is the best word-of-mouth propagator

🔮 What's Next for Trott

The Immediate Future:

As we have a stable product now, we need to market it well
Become profitable
Scale 2K users
Take user feedback, identify their pain points, and then implement

The Dream Features:

Direct import (Bulk) all your Instagram saves
Adding a checklist & notes feature to tick off the places you've travelled, the recipes you've tried, and adding notes next to them.
Amazon and other marketplace integration & Affiliate linking
Other video format and platform support (ex, TikTok, etc.)
Instagram images integration

The Bigger Vision:

Transform how people interact with saved content. No more digital hoarding - just organized, actionable intelligence from content you already loved enough to save.