Portfolio Details
Let's dive deeper into PRD and Product Prototypes!
AI Recipe Generator - Project Information
- Category: Web App
- Ownership: End-to-end (Product, UX, Engineering)
- Project date: June, 2025
- Stack: OpenAI GPT-4o (image generation), Gemini API (recipe creation)
- Github: https://github.com/sahil265wins/recipe-ai-reveal
Project Overview
Problem
High Level Approach
Narrative
- International Student Post-Shopping: Priya opens the AI Recipe Generator, snaps a photo of her shopping bag, and instantly gets several recipe ideas using those ingredients—no need to call her mom.
- Preventing Food Waste: Raj notices leftover veggies. The app suggests a stir-fry recipe, letting him use everything up before it spoils.
- Fast Meal for Busy Professional: Anna uploads a photo of her fridge contents. The app recommends a 15-minute quick meal.
- Independent Cooking Decisions: Users rely on the app for meal inspiration tailored to what’s on hand.
- Future (Grocery Planning): At the store, a user inputs what’s at home and receives a short list of items to buy in order to make a complete meal.
Goals
- Primary Goals
- Maximize recipe generation per user session.
- Reduce decision fatigue and increase confidence.
- Empower new cooks—especially international students.
- Encourage ingredient utilization to reduce food waste.
- Create a simple, trusted daily tool.
- Secondary Goals
- Drive recurring engagement (7-day and 30-day retention).
- High photo-to-recipe conversion rates.
- Technical groundwork for advanced personalization (dietary filters, journaling) in future versions.
Metrics
Metric | Definition | Goal/Guardrail |
---|---|---|
North Star | Recipes generated per user | Higher is better |
DAUs | Daily active users | Minimum week-over-week retention |
Retention Rate | % active users over 7/30 days | Healthy trend; flag <15% 7-day |
Session Time | Avg. session time | 2+ minutes/session |
Photo→Recipe Conversion | % photo uploads yielding a recipe | >80% MVP, >90% target (less than 60% unacceptable) |
Future: Share Rate | % of sessions with recipe sharing | Track post-launch |
Impact Sizing Model (Assumptions made)
- Target: 3,000 students, 25% signup ⇒ 750 users
- Avg 3 recipes/user/week = 2,250 recipes/week
- DAU target: 150
- Monthly: 9,000+ recipes
- Potential revenue: 75 premium users × $3 = $225/month
Non-goals
- Dietary or Allergy Filters in V1
- Recipe History/Journaling
- Grocery Planning and Recommendations
- Native Mobile App
- Elaborate Onboarding or Personalization
Solution Alignment
- V1 Includes: Photo upload, AI ingredient recognition, recipe generation, manual input, mobile web interface
- V1 Excludes: Filters, planners, journaling, history, native apps
Key Features
- MVP: Image Upload, AI Detection, Recipe Generation, Mobile Experience
- Post-MVP: Dietary filters, Sharing, History, Filtering, Planning, Native Apps
Future Considerations
- Personalization: Diet, regional cuisines
- Community Features
- Meal Planning
- Notifications
- Standalone Mobile Apps
Key Flows
- Image-to-Recipe: Upload → Detect → Confirm → Generate → View
- Edge Cases: Prompt manual input for failures or ambiguities
Key Logic
- AI recognition with fallback
- Recipe match by simplicity and completeness
- Friendly error handling
Launch Plan
- Phase 1: Soft Rollout & Testing
- Phase 2: Feedback Iterations
- Phase 3: Success Monitoring & Future Readiness
Major Risks & Contingencies
- Recognition inaccuracy: fallback and model improvements
- Adoption issues: direct outreach and iteration
- Technical issues: ongoing mobile testing, API controlling
Key Milestones
Milestone | Timeline | Owner(s) | Status |
---|---|---|---|
Core MVP Feature Built | ✅ Complete | Dev Team | Done |
End-to-End QA & UX Polish | Week 1 | All | In progress |
Soft Launch/Test to First Users | Week 2 | PM/Dev | Upcoming |
Data Collection & Feedback Sprint | Week 3-4 | PM | Upcoming |
Post-MVP Roadmap Defined | Week 4+ | PM/Team | Upcoming |