GreenPrint 🌱

AI-Powered Sustainability Intelligence for Product Ideas

Transforms hours of manual sustainability research into a ~1-minute AI-generated Life Cycle Assessment — scoring product ideas across 4 frameworks, 6 lifecycle stages, and delivering 5+ cited improvement recommendations per analysis.

🚀 Overview

Sustainability teams and product designers spend 10–40 hours manually assembling Life Cycle Assessments from fragmented academic sources and emissions databases. GreenPrint eliminates that bottleneck.

Architected a full-stack AI product that accepts a product idea as plain text and, in ~1 minute, returns an exhaustive LCA powered by Gemini 3.1 Pro with real-time Google Search grounding — mapping environmental impact across 6 lifecycle stages (raw materials → end-of-life), scoring alignment with 5–7 UN Sustainable Development Goals, rating circularity across the 6Rs framework, and producing 3–5 specific, modular improvement recommendations with estimated impact. All analyses are persisted to Firestore for authenticated users and exportable as branded PDF reports — turning a research-intensive workflow into a self-service, repeatable product tool.

⚙️ Key Features

Runs full LCA analysis in ~1 minute — Engineered a Gemini 3.1 Pro pipeline with Search grounding that replaces hours of manual desk research, returning carbon estimates across 6 lifecycle stages with uncertainty ranges and in-text citations.
Scores across 4 sustainability frameworks simultaneously — Integrated Carbon Footprint (LCA), UN SDGs, 6Rs of Circularity (Refuse / Reduce / Reuse / Repair / Repurpose / Recycle), and the 3 Pillars (Environmental / Social / Economic) into a single structured JSON schema, enabling multi-dimensional product evaluation in one pass.
Compares product ideas side-by-side — Shipped a comparison mode that analyzes multiple product ideas and renders a ComparisonDashboard, enabling designers to evaluate eco-alternatives against each other before committing to a direction.
Exports PDF reports — Integrated html2pdf.js to capture the full analysis dashboard as a branded, shareable PDF report — enabling the output to travel outside the tool into stakeholder decks and vendor reviews.
Persists history to Firestore with localStorage fallback — Implemented a dual-storage abstraction (storage.ts) that writes analyses to Firestore for authenticated users and localStorage for anonymous sessions, keeping history accessible across page reloads with zero friction.
Authenticated sessions via Google OAuth — Integrated Firebase Auth with Google SSO, Firestore user document creation on first sign-in, and security rules scoped per userId — ensuring multi-user safety without a custom backend.
Dark / light mode + first-visit onboarding — Shipped a ThemeToggle and Onboarding overlay (shown once via localStorage flag) to reduce time-to-first-analysis for new users.

🛠️ Tech Stack

Frontend

AI / ML

Infrastructure / Backend

🗺️ Project Layout

Expand file tree

GreenPrint/
├── src/
│   ├── App.tsx                        # Root component: view routing, analysis orchestration
│   ├── main.tsx                       # React entry point
│   ├── index.css                      # Global styles, CSS variables
│   │
│   ├── components/
│   │   ├── IdeaForm.tsx               # Multi-field product idea submission form
│   │   ├── AnalysisDashboard.tsx      # Full single-product LCA results & charts
│   │   ├── ComparisonDashboard.tsx    # Side-by-side multi-product comparison view
│   │   ├── HistoryDashboard.tsx       # Past analyses browser (load, compare, delete)
│   │   ├── AuthProvider.tsx           # Firebase Auth context + Firestore user management
│   │   ├── LoginModal.tsx             # Google SSO sign-in modal
│   │   ├── LoadingState.tsx           # Step-by-step analysis progress indicator
│   │   ├── Glossary.tsx               # Sustainability terminology reference panel
│   │   ├── Onboarding.tsx             # First-visit guided tour overlay
│   │   ├── ThemeToggle.tsx            # Dark / light mode switcher
│   │   └── ui/
│   │       ├── Accordion.tsx          # Collapsible section component
│   │       ├── Badge.tsx              # Label/tag with color variants
│   │       ├── Button.tsx             # Primary/secondary/ghost button variants
│   │       ├── Card.tsx               # Surface container with border & shadow
│   │       ├── Input.tsx              # Controlled text input
│   │       ├── Textarea.tsx           # Controlled multi-line input
│   │       └── Tooltip.tsx            # Hover tooltip wrapper
│   │
│   ├── lib/
│   │   ├── gemini.ts                  # Gemini AI client, LCA prompt, response parser
│   │   ├── firebase.ts                # Firebase app initialization
│   │   ├── firebaseTypes.ts           # Firestore operation type enums
│   │   ├── storage.ts                 # Firestore / localStorage abstraction layer
│   │   ├── pdf.ts                     # html2pdf.js PDF export wrapper
│   │   └── utils.ts                   # General-purpose helpers (classnames, etc.)
│   │
│   └── types/
│       └── index.ts                   # TypeScript interfaces for all domain models
│
├── firestore.rules                    # Security rules: per-user data isolation
├── firebase-blueprint.json            # Firebase project blueprint
├── firebase-applet-config.json        # AI Studio applet config
├── .env.example                       # Required environment variable template
├── vite.config.ts                     # Vite config with env var injection
├── tsconfig.json
└── package.json

🔐 Auth & Login Flow

sequenceDiagram
    participant U as User
    participant A as App
    participant FB as Firebase Auth
    participant FS as Firestore

    U->>A: Click "Sign In"
    A->>FB: signInWithPopup(GoogleAuthProvider)
    FB-->>U: Google OAuth consent screen
    U->>FB: Approves
    FB-->>A: User credential returned
    A->>FS: getDoc(users/{uid})
    alt First sign-in
        FS-->>A: Document does not exist
        A->>FS: setDoc(users/{uid}, { name, email, createdAt })
    else Returning user
        FS-->>A: User document returned
    end
    A->>A: setUser(currentUser) — hydrates AuthContext
    A-->>U: UI updates: avatar shown, history synced to Firestore

Anonymous session: Analysis history is written to localStorage via the storage abstraction. On sign-in, the Firestore layer takes over — analyses submitted while authenticated are scoped to the user via Firestore security rules (request.auth.uid == userId).

🧩 Application State — AppContext

Global state is managed in App.tsx via React hooks and passed down through AuthContext from AuthProvider.

State Slice	Type	Purpose
`ideas`	`ProductIdea[]`	Current analysis subjects (1 for single, N for comparison)
`view`	`'form' \| 'analysis' \| 'comparison' \| 'history'`	Active view routing
`isAnalyzing`	`boolean`	Triggers `LoadingState` overlay, disables navigation
`loadingMsg`	`string`	Step-by-step progress message from `analyzeProductIdea()`
`historyItems`	`ProductIdea[]`	Past analyses loaded from storage on `view === 'history'`
`showGlossary`	`boolean`	Controls Glossary panel visibility
`showOnboarding`	`boolean`	Controls Onboarding overlay (shown once via localStorage flag)
`showLogin`	`boolean`	Controls LoginModal visibility
`user`	`Firebase User \| null`	Auth state from `AuthProvider` context

Data flow: IdeaForm → startAnalysis() → analyzeProductIdea() (Gemini API) → addOrUpdateIdea() (Firestore / localStorage) → setIdeas() → view switches to analysis or comparison.

📋 Pages & Views

View	Trigger	Purpose	Key Components
`form`	Default / "Compare Ideas" click	Product idea input	`IdeaForm`
`analysis`	Single-idea submission	Full LCA results, charts, PDF export	`AnalysisDashboard`
`comparison`	Multi-idea submission	Side-by-side framework comparison	`ComparisonDashboard`
`history`	"History" button	Browse, reload, or compare past analyses	`HistoryDashboard`

🧩 Components

Feature Components

Component	Purpose
`IdeaForm`	Collects product name, description, category, materials, production location, target market, lifespan, and distribution channel
`AnalysisDashboard`	Renders Overall Score, Verdict, Comparative Analysis, Carbon chart (Recharts), SDG badges, 6Rs radar, 3 Pillars breakdown, Improvements, Citations
`ComparisonDashboard`	Renders side-by-side scores for all 4 frameworks across multiple ideas
`HistoryDashboard`	Lists past analyses with load, multi-select compare, and delete actions
`AuthProvider`	Firebase Auth context — exposes `user`, `signIn`, `logOut`, `updateProfileDetails`
`LoginModal`	Google SSO sign-in sheet with Firebase `signInWithPopup`
`LoadingState`	Animated progress indicator with live step messages from the Gemini pipeline
`Glossary`	Sustainability term reference panel (LCA, SDGs, 6Rs, 3 Pillars)
`Onboarding`	First-visit guided tour — shown once, dismissed to localStorage
`ThemeToggle`	Dark / light mode toggle wired to Tailwind CSS class strategy

UI Primitives (src/components/ui/)

Accordion · Badge · Button · Card · Input · Textarea · Tooltip

🧠 NLP & AI Routing

GreenPrint's analysis pipeline is built around a single, highly-engineered prompt in src/lib/gemini.ts.

Model: gemini-3.1-pro-preview Tools: { googleSearch: {} } — enables real-time web grounding for up-to-date emissions factors, SDG targets, and circularity research Temperature: 0.2 — set deliberately low to maximize analytical precision and reduce hallucination Response format: application/json with strict schema enforcement

Prompt design principles:

Assigns a senior sustainability analyst persona with explicit LCA methodology instructions
Requires comparison against the most common conventional alternative (e.g., bamboo vs. plastic toothbrush across all 6 lifecycle stages)
Enforces in-text citation format [1], [2] matched to a citations array in the JSON response
Handles digital products with adapted lifecycle stages (server energy, data transfer, device load)
Instructs the model to state estimation methodology explicitly when hard data is unavailable

Prompt → Response flow:

flowchart LR
    A[User submits ProductIdea] --> B[analyzeProductIdea]
    B --> C{Build prompt\nfrom idea fields}
    C --> D[Gemini 3.1 Pro\n+ Search Grounding]
    D --> E[Web search:\nemissions factors,\nSDG targets,\ncircularity reports]
    E --> F[JSON response:\noverallScore, carbon,\nsdgs, sixRs,\nthreePillars,\nimprovements, citations]
    F --> G{JSON.parse}
    G -->|Valid| H[SustainabilityAnalysis object]
    G -->|Invalid| I[Error thrown\nto UI alert]
    H --> J[Rendered in\nAnalysisDashboard]

Output schema (TypeScript):

interface SustainabilityAnalysis {
  overallScore: number;           // 0–100
  verdict: string;                // Multi-paragraph summary
  comparativeAnalysis?: string;   // vs. conventional alternative
  carbon: CarbonFootprint;        // 6 lifecycle stages + totalKg + uncertainty
  sdgs: SDGs;                     // 5–7 goals + target codes + explanation
  sixRs: SixRs;                   // 6 sub-scores (0–10) + reasons
  threePillars: ThreePillars;     // Environmental / Social / Economic scores
  improvements: Improvement[];    // 3–5 modular recommendations + impact estimate
  citations: Citation[];          // 5+ web sources with url + snippet
}

Fallback handling: JSON.parse errors are caught and rethrown to the UI as an alert(). The storage layer is only written to on successful parse — invalid responses are never persisted.

🔧 Hooks & Utilities

File	Purpose
`src/lib/gemini.ts`	Initializes `GoogleGenAI` singleton, builds LCA prompt from `ProductIdea` fields, calls Gemini API, parses and validates JSON response
`src/lib/storage.ts`	`loadHistory()` / `addOrUpdateIdea()` / `deleteIdea()` — routes to Firestore (authenticated) or localStorage (anonymous) transparently
`src/lib/firebase.ts`	Initializes Firebase app, `auth`, and `db` instances from environment config
`src/lib/pdf.ts`	Wraps `html2pdf.js` to capture a DOM element by ID and trigger browser download
`src/lib/utils.ts`	`cn()` utility merging Tailwind classes via `clsx` + `tailwind-merge`
`src/components/AuthProvider.tsx`	Exports `useAuth()` hook — provides `user`, `signIn`, `logOut`, `updateProfileDetails` to any component in the tree

🚀 Getting Started

Prerequisites: Node.js 18+, a Google Gemini API key, and a Firebase project with Auth + Firestore enabled.

# 1. Clone the repository
git clone https://github.com/yatinbhalla/EcoLens.git
cd EcoLens

# 2. Install dependencies
npm install

# 3. Configure environment variables
cp .env.example .env.local

Required environment variables (.env.local)

GEMINI_API_KEY=your_gemini_api_key_here

# Firebase config (from Project Settings > General > Your apps)
VITE_FIREBASE_API_KEY=
VITE_FIREBASE_AUTH_DOMAIN=
VITE_FIREBASE_PROJECT_ID=
VITE_FIREBASE_STORAGE_BUCKET=
VITE_FIREBASE_MESSAGING_SENDER_ID=
VITE_FIREBASE_APP_ID=

# 4. Start the development server
npm run dev
# → http://localhost:3000

# 5. Build for production
npm run build

# 6. Preview production build
npm run preview

Or try it live → GreenPrint on AI Studio (no setup required)

🚧 Known Limitations

API key is self-managed — there is no managed backend proxy; the Gemini key is injected at build time via Vite, which exposes it in the client bundle. For production, route calls through a serverless function.
Analysis error handling is alert-based — failed Gemini calls currently surface as window.alert(). A v2 priority is inline error states within the dashboard.
No input validation before API call — the form submits any text to Gemini without pre-processing. Garbage input produces low-quality output rather than a clear validation error.
PDF export uses DOM capture — html2pdf.js captures the rendered DOM, which can produce layout artifacts for long analyses or on non-standard screen widths.
localStorage analyses don't migrate on sign-in — analyses made anonymously are not automatically merged into Firestore when the user signs in.
Single Gemini model — gemini-3.1-pro-preview is hardcoded; no fallback to a faster/cheaper model for lighter queries.

🤝 Contributing

GreenPrint is an open product-in-progress and I genuinely welcome collaborators. Whether you've spotted a prompt engineering improvement, have ideas for new sustainability frameworks (LEED? Scope 3?), or want to tackle one of the Known Limitations above — please open an issue or PR.

Found a bug? Open an Issue with steps to reproduce and your browser/OS.
Have a feature idea? Describe the user problem you're solving, not just the solution — product critiques especially welcome.
Want to contribute code? Fork the repo, create a feature branch, and open a PR against main. I'll review within 48 hours.

Author

Yatin Bhalla · Product Manager & AI Product Builder

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
src		src
.env.example		.env.example
.gitignore		.gitignore
README.md		README.md
eslint.config.js		eslint.config.js
firebase-applet-config.json		firebase-applet-config.json
firebase-blueprint.json		firebase-blueprint.json
firestore.rules		firestore.rules
index.html		index.html
metadata.json		metadata.json
package-lock.json		package-lock.json
package.json		package.json
tsconfig.json		tsconfig.json
vite.config.ts		vite.config.ts

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

GreenPrint 🌱

AI-Powered Sustainability Intelligence for Product Ideas

📋 Table of Contents

🚀 Overview

⚙️ Key Features

🛠️ Tech Stack

🗺️ Project Layout

🔐 Auth & Login Flow

🧩 Application State — AppContext

📋 Pages & Views

🧩 Components

🧠 NLP & AI Routing

🔧 Hooks & Utilities

🚀 Getting Started

🚧 Known Limitations

🤝 Contributing

Author

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

GreenPrint 🌱

AI-Powered Sustainability Intelligence for Product Ideas

📋 Table of Contents

🚀 Overview

⚙️ Key Features

🛠️ Tech Stack

🗺️ Project Layout

🔐 Auth & Login Flow

🧩 Application State — AppContext

📋 Pages & Views

🧩 Components

🧠 NLP & AI Routing

🔧 Hooks & Utilities

🚀 Getting Started

🚧 Known Limitations

🤝 Contributing

Author

About

Topics

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages