NestCash

Problem

NestCash is a real estate lead generation business targeting motivated home sellers across 10 states.

Target sellers

• Pre-foreclosure and behind on mortgage
• Divorce and inherited property
• As-is homes that need to sell fast

The challenge

Compete in a keyword space dominated by national players by publishing faster and smarter at scale.

What I Built

Content at scale

• 70 city landing pages across 10 states with local market data (median price, days on market, cash transaction %, foreclosure rate)
• 10 state hub pages plus 4 situational hubs (divorce, foreclosure, as-is, inherited), each 1,800+ words with 12 to 14 FAQs
• 273+ blog articles targeting city keywords and situational search intent

Automation

• Dual content pipeline: city pipeline nightly, situational pipeline daily, both via GitHub Actions
• LLM generation with Claude, Gemini, and GPT-4o fallback
• Firecrawl competitor scraping and a 7-invariant audit before publish

Ops and delivery

• HubSpot CRM and Twilio SMS for lead capture and alerts
• Deployed on Railway via Docker with auto-deploy from main

What Made It Interesting

This is less a website than a content production system wearing a website's clothes. The pipeline is not just "call LLM, save file."

• Brief generator pulls competitor data and market context
• Cluster-aware deduplication reads up to 8 similar articles before generating
• Author rotation and automatic cross-link injection
• Audit step enforces structural invariants before anything ships
• Runs end to end without manual intervention

Pipeline

1. Brief. Pull competitor data via Firecrawl, attach market context and state legal info.
2. Generate. Cluster-aware dedup reads up to 8 similar articles, then Claude generates with author rotation.
3. Audit. 7 structural invariants enforced before anything ships.
4. Inject links. Automatic cross-link injection with bracket-balance safety checks.
5. Publish. Commit to main, Railway auto-deploy, Search Console ping.

Stack