FitFork is a next-generation, RAG-powered culinary assistant designed to bridge the gap between metabolic requirements and professional recipe execution. It transforms complex user metrics into actionable, dietary-compliant meal plans using state-of-the-art vector retrieval and large language models.
For a deep dive into how FitFork works and how to set it up, please refer to our detailed guides:
Traditional meal planners rely on rigid templates. FitFork treats nutrition as a dynamic data problem. By analyzing height, weight, activity levels, and fitness goals (Mifflin-St Jeor accuracy), the system generates a unique caloric and macronutrient fingerprint for every user.
FitFork features a premium dark botanical aesthetic. Designed for the modern kitchen, the high-contrast "Deep Olive & Cream" palette ensures readability in low-light environments while maintaining a grounded, sophisticated feel.
Our solution is built on a high-concurrency, biometric-aware architecture that unifies nutrition science with modern retrieval-augmented generation in a high-contrast Nano Banana aesthetic.
graph TD
%% User Interaction
U[User] -->|Biometrics / Preferences| F[React Frontend]
F -->|Secure JWT Auth| B[FastAPI Backend]
%% Metabolic Engine
subgraph "Metabolic Engine (Nano Banana Core)"
B -->|Mifflin-St Jeor| M[BMR/TDEE Processor]
M -->|Caloric Envelope| P[Prompt Orchestrator]
end
%% Unified RAG Pipeline
subgraph "Unified RAG Store"
P -->|Semantic Query| DB[(MongoDB Vector Store)]
DB -->|Biometric-Filtered Recipes| P
end
%% Intelligence Layer
P -->|Seeded Context| AI[Gemini 2.0 Flash]
AI -->|JSON Meal Plan| B
%% Calendar Sync
B -->|OAuth 2.0| GCal[Google Calendar API]
GCal -->|Sync Events| U
%% Styling (Nano Banana High-Contrast)
style U fill:#f9f9f9,stroke:#FFD700,stroke-width:3px
style F fill:#333,stroke:#FFD700,stroke-width:2px,color:#fff
style B fill:#333,stroke:#FFD700,stroke-width:2px,color:#fff
style DB fill:#333,stroke:#FFD700,stroke-width:2px,color:#fff
style AI fill:#333,stroke:#FFD700,stroke-width:2px,color:#fff
style GCal fill:#333,stroke:#FFD700,stroke-width:2px,color:#fff
style M fill:#FFD700,stroke:#333,color:#333
style P fill:#FFD700,stroke:#333,color:#333
- FastAPI: Asynchronous, high-performance API layer.
- MongoDB: Unified store for recipes, user profiles, authentication, and chat history.
- Google GenAI SDK: Native interface for Gemini 2.5 models.
- Pydantic: Strict data validation for complex nutritional schemas.
- Vite + React: Modern, lightning-fast rendering engine.
- Shadcn UI: For premium, accessible component architecture.
- Framer Motion: Subtle micro-animations for an alive, interactive interface.
- Tailwind CSS: Custom botanical tokens for a unified design system.
- Python 3.10+
- Node.js 18+
- MongoDB instance (Local or Atlas)
- API Keys: OpenRouter, Scaledown (Optional)
cd backend
python -m venv venv
source venv/bin/activate # On Windows: .\venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env # Configure your environment variables
uvicorn app.main:app --reloadcd frontend
npm install
npm run dev| Endpoint | Method | Purpose |
|---|---|---|
/auth/signup |
POST | Resident registration with hashed credentials. |
/user/nutrition |
POST | Calculate BMR/TDEE and persist profile. |
/search |
POST | Personalized RAG recipe retrieval. |
/meal-plan |
POST | Generate full interactive calendar plan. |
/health |
GET | System integrity check. |
- Phase 4: In-app Grocery List generator based on weekly recipes.
- Phase 5: Real-time pantry tracking via image recognition.
- Phase 6: Integration with wearable health data (Apple Health/Google Fit).
This project is licensed under the MIT License - see the LICENSE file for details.
FitFork is built to empower individuals to take control of their nutrition without sacrificing the joy of professional-grade cooking. 🌿🍔