Water Tracker AI Agent

A full-stack hydration assistant that logs daily water intake, stores history, and uses an LLM to generate personalized feedback in real time.

Why this project stands out

• Combines a clean Streamlit UX with an AI feedback loop so the user gets instant, actionable insights.
• Persists intake history in SQLite and visualizes trends over time with built-in charts.
• Exposes a FastAPI backend for programmatic logging and retrieval, enabling future mobile or wearable integrations.

Features

• Log water intake per user and receive AI-generated hydration guidance.
• View historical intake data as a table and line chart.
• FastAPI endpoints for logging and history lookup.
• Simple local setup with .env-based OpenAI API configuration.

Tech stack

• Python
• Streamlit UI
• FastAPI + Pydantic
• SQLite
• LangChain + OpenAI

Architecture

• dashboard.py handles the Streamlit user interface.
• src/api.py exposes REST endpoints for intake logging and history.
• src/database.py writes and reads from water_tracker.db.
• src/agent.py invokes the LLM for hydration analysis.

Quick start

1) Install dependencies

pip install -r requirements.txt

2) Configure environment

Create a .env file with your OpenAI key:

OPENAI_API_KEY=your_key_here

3) Run the Streamlit dashboard

streamlit run dashboard.py

4) Run the FastAPI service

uvicorn src.api:app --reload

API endpoints

• POST /log-intake
  • Body: { "user_id": "user_123", "intake_ml": 1000 }
• GET /history/{user_id}

Project structure

.
+-- dashboard.py
+-- src
�   +-- agent.py
�   +-- api.py
�   +-- database.py
�   +-- logger.py
+-- requirements.txt
+-- water_tracker.db

Results

A simple, fast, and visually clean hydration tracker that blends practical data logging with AI-driven coaching.