AI Restaurant Menu Research SaaS

I am building a Software-as-a-Service platform that lets restaurant owners instantly research their competitors’ menus and pricing. The very first release is focused solely on one task: gathering live menu items with their associated prices from restaurant websites, structuring that data, and presenting clear side-by-side comparisons.

Here is what I need developed:

• A web crawler or scraper that reliably pulls menu item names and prices directly from publicly available restaurant websites, coping with common menu layouts (HTML tables, menu plugins, simple text blocks).
• An AI-driven parser that cleans the raw HTML, detects currency and serving sizes, and normalises the results so similar dishes can be compared regardless of formatting differences.
• A secure, multi-tenant backend (preferably Python + FastAPI or Node.js) that stores the scraped data and keeps it up to date on a configurable schedule.
• A lightweight dashboard where users can type in competitor URLs, trigger a fresh scan, and instantly see a comparison table highlighting price gaps, average dish cost, and top most expensive or cheapest items.
• Clear output endpoints (REST or GraphQL) for future features such as AI assisted menu creation.

Acceptance criteria
1. When given a valid restaurant website URL the system returns a JSON array of dishes with name, price, currency, and the page URL where each was found.
2. Price accuracy must be within ±2 % of the value shown on the live site.
3. The dashboard loads the comparison table in under three seconds for up to 10 competitors.
4. All scraping modules obey robots.txt and rate-limit rules.

If you have proven experience building web scrapers, cleaning semi-structured data with NLP, and packaging that into a SaaS dashboard, I’d love to see a short outline of your proposed tech stack, milestone plan, and any relevant demos.