I need an end-to-end, low-cost MVP that pulls public car and real-estate ads from Dubizzle, ContactCars, Property Finder, Hatla2ee, and Aqarmap, stores them in my own database, and presents them through both a web interface and a lightweight mobile app. Scope of work • Backend with one shared API that feeds a web frontend (including an admin area) and a simple Flutter or React Native app. • A data-collection layer able to crawl or consume the open listings on the five marketplaces above, capturing title, price, photos, specs, location, and description, then de-duplicating and saving them in my tables. • Basic search plus category and attribute filters (price, location, model, etc.) on both web and mobile. • Admin panel functions: create/edit categories, approve or remove individual listings, and view real-time scraping status or error logs. • Clean, minimalist UI; no heavy graphics or animations. • Deployment on a budget-friendly host—feel free to suggest the most economical stack that still scales. Deliverables 1. Source code for backend, web frontend, and mobile app. 2. Database schema and seed data from initial scrape. 3. Admin credentials with full access. 4. Simple README covering setup, deployment, and how to add new marketplaces later. Acceptance criteria • Listings from all five named sources appear correctly in both web and app within one minute of ingestion. • Search and filters return accurate results in under two seconds on a modest hosting plan. • Admin actions (approve, remove) propagate to both channels instantly without reload. • No marketplace is rate-limited or blocked during a 48-hour continuous scrape test. Tech choices are open as long as they stay cost-effective and future-proof; common stacks like Node.js + PostgreSQL or Django + Postgres are perfectly fine. Let me know if any part of the flow needs clarification before starting.