Skip to content

bhineswaveformer6/squire-cre-scraper

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Squire — Live CRE Intelligence Agent

"Batman doesn't wait for the deal to come to him."

Built by Brandon Hines [Ψ̂-001] · CortexChain, Inc. × Waveform Tech LLC


What It Is

Squire is a multi-source commercial real estate data pipeline that:

  1. Scrapes live listings from CoStar, LoopNet, and Crexi
  2. Qualifies assets through cap rate + DSCR gates
  3. Injects trophy-grade targets into the AgentSeeker / Forge system

One command. Three sources. Only the deals that pass the Batman gate come out the other side.


Architecture

SquireScraper
    ├── Source 1: Crexi (JSON API + HTML fallback via ScraperAPI)
    ├── Source 2: LoopNet (Playwright headless browser)
    ├── Source 3: CoStar (authenticated session)
    ├── DataNormalizer → unified CREListing schema
    ├── QualificationFilter → cap rate ≥ 5%, DSCR ≥ 1.25
    └── Output → JSON export / AgentSeeker inject

Install

pip install playwright httpx beautifulsoup4 scraperapi-sdk python-dotenv rich tenacity
playwright install chromium

Environment Variables

SCRAPERAPI_KEY=your_scraperapi_key
COSTAR_USERNAME=your_costar_email
COSTAR_PASSWORD=your_costar_password
CREXI_API_KEY=your_crexi_api_key
MIN_CAP_RATE=0.05
MIN_DSCR=1.25
TARGET_MARKETS=New York City,Los Angeles,Miami,Dallas,Chicago

Run

# Single market scan
python squire_scraper.py --market "Miami" --asset-type office

# Full pipeline (all TARGET_MARKETS)
python squire_scraper.py --all

# Output to JSON
python squire_scraper.py --all --output results.json

Output Schema

{
  "listing_id": "LST-A1B2C3D4",
  "source": "Crexi",
  "title": "Class A Office Tower",
  "asset_type": "office",
  "market": "Miami",
  "address": "1000 Brickell Ave, Miami, FL 33131",
  "asking_price": 12500000,
  "noi": 750000,
  "cap_rate": 0.06,
  "sqft": 45000,
  "year_built": 2018,
  "url": "https://www.crexi.com/properties/...",
  "qualified": true,
  "qualify_score": 8.31,
  "scraped_at": "2026-03-28T01:30:00"
}

Qualification Gate (Batman Filter)

  • cap_rate >= MIN_CAP_RATE (default 5%)
  • estimated_dscr >= MIN_DSCR (default 1.25x, derived from cap rate / 6.5% debt constant)
  • asking_price > 0

Only assets passing all three gates are marked qualified: true.


Part of the Batman Robb Report Stack

Squire (this repo)     → live CRE data pipeline
AgentSeeker            → founder/investor matching
Forge                  → deal structuring engine
Batman Robb Report     → sovereign command surface

Built with Perplexity Computer · Powered by NVIDIA NIM

About

Batman Robb Report — Live CRE Agent: CoStar | LoopNet | Crexi multi-source scraper + qualification engine

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages