Smart-Appetite-Manager

Smart Appetite Manager is an AI-powered platform designed to help you manage your entire grocery cycle. It tracks your inventory, suggests recipes based on what you already have, and finds the best grocery deals in real time. Powered by a team of specialized AI agents built on Solace Agent Mesh framework, it helps you save time, reduce food waste, and shop smarter every day.

Smart Appetite Manager Setup Instructions

Use the App/ folder as the project root (not repository root).

CLI Install + Run (Recommended)

cd App

# 1) Create env file
cp .env.example .env

Edit App/.env and set at least:

LLM_SERVICE_API_KEY=...
LLM_SERVICE_ENDPOINT=https://api.cerebras.ai/v1
LLM_SERVICE_PLANNING_MODEL_NAME=openai/gpt-oss-120b
LLM_SERVICE_GENERAL_MODEL_NAME=openai/gpt-oss-120b
SOLACE_DEV_MODE=true

For this fork, also add:

SPOONACULAR_API_KEY=...        # used by recipe tools (App/src/recipe_agent/mealdb_tools.py:17)
SERPAPI_KEY=...                # used by shopper agent (App/configs/agents/shopper.yaml:70)
INVENTORY_MANAGER_DB_NAME=inventory.db   # used by inventory agent (App/configs/agents/inventory-manager.yaml:52)

Inventory DB bootstrap (automatic): No manual SQL step is required. The inventory tool now runs schema bootstrap on every call and creates/updates the inventory table if needed.

Install dependencies and run:

uv sync
uv run sam run configs/

Open: http://localhost:8000

Inventory web interface (custom page with REST read + SAM write):

• Open App/web/inventory-gateway-ui.html
• Full guide: App/docs/inventory-web-interface.md
• Start inventory API first:

cd App
./.venv/bin/python -m uvicorn src.inventory_api.app:app --host 0.0.0.0 --port 8001 --reload

If uv is missing on macOS:

brew install uv

Docker (Recommended for full stack)

The repo ships a docker-compose.yml that starts all four services together:

Service	Port	Description
sam-app	8000	Solace Agent Mesh (AI orchestration)
inventory-api	8001	Inventory REST API
ocr-service	8002	Barcode / OCR service
web	5173	React frontend

Prerequisites: Docker Desktop (includes Compose v2).

# 1. Copy and fill in env vars (repo root)
cp App/.env.example App/.env
# edit App/.env — set LLM keys, SPOONACULAR_API_KEY, SERPAPI_KEY
# INVENTORY_MANAGER_DB_NAME is set automatically to /app/data/inventory.db

# 2. Build and start
docker compose up --build -d

# 3. Open the app
open http://localhost:5173

The inventory database is stored in a named Docker volume (inventory_data) mounted at /app/data inside the containers. It persists across restarts and is created automatically on first run — no manual SQL step required.

Logs / stop:

docker compose logs -f          # stream all services
docker compose logs inventory-api   # single service
docker compose down             # stop and remove containers (volume kept)
docker compose down -v          # stop and remove containers + volume (wipes DB)