MagicBricks Scraper — India Real Estate Listings

Scrape live India property listings from MagicBricks — price, area, locality, BHK and property type — for sale or rent, across Mumbai, Bangalore, Delhi NCR, Pune, Hyderabad, Chennai and 100+ other Indian cities.

Market data for your research. Not investment, valuation or financial advice.

⏰ Run this as a daily price/listing watch (recommended)

Prices and inventory move constantly. Schedule a daily sweep (0 7 * * *) and track changes over time.

📊 What you get

Clean JSON, one record per listing. The real field names are:

• property_id — MagicBricks listing id
• title — listing headline as posted (e.g. "2 BHK Flat for Sale in Anjanapura Bangalore")
• city — city searched
• locality — locality, from the card or derived from the listing title
• builder — project / developer name where MagicBricks exposes one (empty when it does not)
• bhk — bedroom count
• property_type — e.g. Multistorey Apartment, Villa, Residential Plot
• area_sqft — advertised area
• price_inr — price in rupees (total for sale, per month for rent)
• price_display — Indian-format price string, e.g. "₹ 97.2 Lac"
• price_per_sqft — derived ₹/sqft
• listing_type — sale or rent
• property_age, furnishing, posting_date — populated when MagicBricks' JSON path is available; often empty on the HTML search-page path
• contact_url — the MagicBricks listing URL
• source_url, data_source, scraped_at — provenance

Fields MagicBricks does not publish on a given listing are returned empty. Nothing is invented.

💵 Pricing

Pay-per-event — you pay only for listings that land in your dataset:

• Actor start: $0.00005 per run (per GB of run memory)
• Result: $0.05 per listing — 20 listings per $1

10 listings ≈ $0.50 · 100 listings ≈ $5.00 · a daily 50-listing watch ≈ $75 / month.

✅ Run outcomes

Every delivered row is a real listing read from magicbricks.com during that run.

• Listings found — run succeeds, you are charged per listing.
• MagicBricks matched nothing (its "no search results" page) — run succeeds with 0 rows and a status message telling you which filter to widen. Nothing is charged.
• MagicBricks blocked, served a holding page, or changed its markup — the run fails with a status message naming the cause. No rows, no charge.

The actor never writes reference, sample, placeholder or status rows into the dataset.

🤖 Use with AI agents

Point Claude, the OpenAI Agents SDK, an n8n flow or any MCP-aware client at it and pull data on demand.

Optional Notion delivery: connect your own Notion MCP connector and the run also writes the listings to a Notion page. The actor never sees your Notion credentials.

🚀 Quick start

{
  "city": "Bangalore",
  "propertyType": "apartment",
  "listingType": "sale",
  "bhk": ["2"],
  "budgetMaxInr": 10000000,
  "maxResults": 10
}

🔗 Related actors

Global property data — listings and prices across seven markets:

Zillow (US) · Redfin (US) · Rightmove (UK) · Boliga (Denmark) · Centaline (Hong Kong) · Singapore URA · Singapore HDB

Agent front doors: Real Estate MCP · Redfin MCP

---

Public web data, assembled for research and monitoring.