How AI helped build India’s first hyper-local carbon tracker in two weeks

Imagine telling someone their carbon footprint equals 14 flights between Mumbai and Delhi. That’s the kind of insight that turns abstract data into real-world impact—and it’s exactly what drove me to build EcoTrack India, a gamified carbon footprint platform designed for India’s unique energy landscape.

Bridging the gap between numbers and real life

India emits about 1.9 tonnes of CO2 per person annually, but that statistic alone does little to change behavior. Most carbon calculators rely on global averages, leaving users confused about how their choices affect the planet. EcoTrack India tackles this by replacing vague figures with concrete comparisons: your yearly footprint might be equivalent to charging a smartphone half a million times or taking 14 round-trip flights between major cities.

Why traditional carbon calculators miss the mark

After testing half a dozen platforms, I noticed consistent flaws that made them ineffective for Indian users:

• Overgeneralized averages: National electricity emission factors ignore regional differences. For example, coal-heavy Jharkhand (0.95 kg CO2/kWh) produces nearly three times more emissions per unit than hydro-reliant Himachal Pradesh (0.30 kg CO2/kWh).

• Static results: Most tools simply display a number with no follow-up action, leaving users with no clear path to reduce their impact.

• Western-centric design: Features like LPG cylinders, CNG auto-rickshaws, and train-heavy commutes were absent, making the tools irrelevant for local lifestyles.

• No retention strategy: Users calculated their footprint once, felt momentarily guilty, and never returned.

How Google Antigravity transformed development speed

Before discovering Google Antigravity, my development cycle followed a familiar pattern: sketch an idea, draft wireframes, write code, debug, and repeat—a process that could take weeks for a single feature. With Antigravity, the cycle shifted dramatically:

From weeks to minutes

I outlined a clear prompt to build a multi-step carbon calculator tailored to India’s 29 states, incorporating state-specific electricity grid factors, Scope 1 (direct emissions like petrol and LPG), Scope 2 (electricity), and Scope 3 (food, flights, shopping) categories. The prompt also requested a "Carbon Credit Score" styled like a financial credit score, ranging from 0 to 850, with tiered results.

Within two minutes, Antigravity generated a functional prototype with all state grid factors, Scope breakdowns, and a scoring engine—work that would have taken days of manual coding.

Refining the AI-generated tool: three key fixes

Antigravity isn’t a magic solution—it’s a powerful tool that requires precise direction. Here’s how I improved the initial output:

1. Correcting the scoring formula

The first version underestimated scores significantly. A vegetarian in Karnataka with no car was getting 420/850 when they should have been closer to 720+. I refined the prompt to specify the exact scoring formula:

Score = max(0, 850 - (userTonnes / 10 * 850))

This adjustment ensured scores reflected realistic benchmarks.

2. Personalizing behavioral nudges

Generic advice like "reduce energy consumption" didn’t resonate. I reprompted for context-aware tips:

If the user’s highest emitting Scope is electricity AND they live in a coal-heavy state, show tailored tips like installing rooftop solar or shifting AC usage to off-peak hours.

The result? Advice that felt genuinely personal and actionable.

3. Designing shareable, engaging visual cards

The first Canvas-generated share card looked like a basic receipt. I requested a modern design with:

• A glassmorphism style

• A leaf-patterned background

• The user’s tier emoji prominently displayed

• A QR code placeholder reading "Scan to calculate yours!"

This small tweak made the card something users actually wanted to share on social media.

Five standout features of EcoTrack India

1. State-level grid intelligence

EcoTrack India doesn’t use a single national average. Instead, it accounts for regional electricity mixes, so selecting Maharashtra versus Himachal Pradesh with identical electricity consumption yields different scores—because the underlying carbon reality differs.

2. "Making the invisible visible" engine

After calculating a footprint, users see six relatable equivalents:

• "Your footprint equals charging your phone 487,000 times."

• "Your footprint equals 14 Mumbai–Delhi flights."

• "You’d need 43 trees to offset your annual emissions."

These comparisons transform abstract numbers into tangible impact.

3. Real-time behavioral nudges

The app doesn’t rely on intrusive pop-ups. Instead, contextual toasts appear at the exact moment of decision-making:

• Selecting an omnivore diet triggers: "Shifting to vegetarian could save ~1,000 kg CO2/year—that’s like parking your car for 5 months."

4. Five-step onboarding tour (vanilla JS)

The entire onboarding experience is built with vanilla JavaScript and no external libraries. It uses scrollIntoView() for smooth navigation, CSS z-index stacking for overlays, and localStorage to remember progress—ensuring first-time users never see the tour again.

5. Shareable Carbon Credit Card (Canvas API)

Every user gets a unique, programmatic image generated using HTML5 Canvas. The card includes:

• The user’s exact score and tier

• Their state and city benchmark

• A downloadable PNG ready for social sharing

No Canva templates or third-party tools were used—just pure code.

Behind the data: a meticulously curated database

The foundation of EcoTrack India is its comprehensive emission factor database, which includes:

• 29 Indian states, each with unique grid emission factors

• 5 diet types, sourced from scientific studies on annual emissions

• 8 transport modes, accounting for India-specific options like CNG auto-rickshaws and 3-wheelers

• 8 city benchmarks for peer comparison

• 3 Scope categories, aligned with enterprise carbon accounting standards

Vibe coding vs. traditional development: a paradigm shift

Traditional development feels like being both architect and bricklayer—you design the structure and lay every stone yourself. Vibe coding flips this model: you remain the architect, but a skilled crew handles execution. The bottleneck shifts from "Can I code this?" to "Can I describe this clearly enough?"

This approach demands precision in communication. Writing effective prompts is now as critical as writing clean code. The speed gains are undeniable, but success hinges on clarity and intentionality—mistakes still happen, but they’re caught faster.

EcoTrack India proves that AI can accelerate innovation without sacrificing quality. By focusing on user-centric design and hyper-local data, the platform transforms dry statistics into meaningful action. The next step? Scaling this approach to other regions and refining the AI’s ability to generate even more personalized insights.