AI Personal Trainer

A simple full-stack AI app that acts as a personalized fitness coach. Users sign up, set their training preferences, and chat with an AI trainer that adapts to their goals, experience level, equipment, and preferred coaching tone.

Stack

• Backend: FastAPI
• Frontend: Streamlit
• LLM: Groq (llama-3.1-8b-instant) via LangChain
• Database: SQLite via SQLAlchemy

Project structure

.
├── main.py              # FastAPI backend (auth, preferences, chat endpoints)
├── dbSQLAlchemy.py      # SQLAlchemy models + CRUD helpers
├── app.py               # Streamlit entry: login / signup
├── pages/frontend.py    # Streamlit page: preferences form + chat UI
├── start.sh             # Launch backend + frontend together
├── requirements.txt
└── .env.example

Setup

1. Create and activate a virtual environment

python -m venv venv
source venv/bin/activate        # macOS / Linux
# venv\Scripts\activate         # Windows

2. Install dependencies

pip install -r requirements.txt

3. Configure environment variables

Copy .env.example to .env and fill in your Groq API key:

cp .env.example .env

GROQ_API=your_groq_api_key_here

You can get an API key from console.groq.com.

Running the app

Option A — start.sh (recommended)

bash start.sh

This starts:

• FastAPI on http://localhost:8000
• Streamlit on http://localhost:8081

Open the Streamlit URL in your browser.

Option B — run the two services manually

In one terminal:

uvicorn main:app --host 0.0.0.0 --port 8000 --reload

In another:

streamlit run app.py

API endpoints

Method	Path	Purpose
POST	/signup/	Create a new user
POST	/login/	Look up an existing user
POST	/update_preferences/	Save / update training preferences
POST	/chat/	Send a message to the AI trainer
GET	/chat_history/{username}	Fetch a user's full chat history

How chat works

On every /chat/ call the backend:

1. Loads the user's saved preferences and full chat history from SQLite.
2. Rebuilds LangChain memory from the stored messages.
3. Builds a system prompt that interpolates the user's goal, experience, days_per_week, equipment, and tone.
4. Runs the prompt + history + new message through the Groq LLM.
5. Persists both the user's message and the assistant's reply.

Notes

• The database file my_database.db is created automatically on first run; there are no migrations.
• Authentication is username-only (no passwords) — this project is intended as a demo, not a production deployment.

License

MIT — see LICENSE.