FastMCP-V1

Technology:

Purpose:

The purpose of this project is to be able to find resturants through LLM that are within a mile radius. When people look for food to eat at you usually factor in price. the user can then prompt the llm host of the budget and the output will be a list of resturants which will be saved to a database. The search results will be reused for 60 days. after that it will rerun the search. a menu parser will be ran for each resurtant and data will be saved to a databse. The results will also be good and reused for 60 days. after then a parser will then need to be ran.

Problem:

use LLM to find something to eat produces very limited results instead of seeming the options and possible combinations that is within someeones budget. FastMCP-V1 solves this by surfacing concrete menu options that fit within a specified budget and distance, with reusable data caching.

Built with

🐍 Python ⚡ FastAPI 🧠 FastMCP Framework 🌐 Google Maps API 📬 Gmail SMTP (for alerts/emails) 🌍 MongoDB (data persistence) 🎭 Playwright (JS-rendered menu scraping) 🧾 PDFPlumber (PDF menu parsing)

Agents & Tools

Local-Time 8001 Returns timezone-aware local time. Send-Email 8002 Sends email notifications via Gmail SMTP. Resturant-Agent 8003 Queries Google Maps API for nearby restaurants and distances. MenuScraperAgent 8004 Scrapes menus (HTML, PDF, Square) and builds budget-friendly combinations. Database-Agent 8005 Stores/retrieves restaurant and menu data in MongoDB.

SetUp

include an .env file which contains keys for api and usernames and passwords for connection to your account.

Mongo DB: MONGO_URI: ENTER IN YOUR URL

gmail: EMAIL_ADDRESS: Enter IN YOUR EMAIL ADDRESS EMAIL_PASSWORD: ENTER IN YOUR EMAIL PASSWORD

google maps api key: