Score off-market real estate deals from emails
A pipeline has a Python worker read deal emails, Claude extract fields into fixed JSON, and code score them; the poster is stuck on incomplete data and login-walled sources.
Done withClaude CodeClaude
- What they did
- The poster wired a dedicated Gmail to collect saved-search alerts and wholesaler emails, then had a Python worker split multi-deal emails and send each deal to a low-cost Claude model to extract fixed JSON fields, leaving unknowns null. Plain Python, not the LLM, handled the single-family and multifamily scoring, with paid lookups only after a cheap price check; output went to Supabase, a Streamlit dashboard, email alerts, and a shared Google Doc for coordination.
- How it went
- It worked well enough to centralize incoming deals and score what had data, but many leads stalled because about half arrived without a street address and others lacked unit counts or rents. The poster says those missing fields, plus login-walled sources that block automation, kept deals stuck at a needs-info stage.
- Worth knowing
- Their main catch is cost control and gating: expensive data lookups only run after a cheap initial price check.
Try it yourself with Claude Code
Help me build a Python script that reads real estate deal emails from [mailbox or folder of exported .eml files], uses Claude to extract fields into a fixed JSON schema (address, asking price, units, rents, NOI, seller contact), and uses plain code to score each deal against [my criteria]. When data is missing, mark the field as null and list it in a 'needs follow-up' output rather than guessing. Done means I can run it on a sample of 10 emails and get a ranked CSV of deals with scores and missing-data flags.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: Reddit · Undated
Related use cases
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Build an invoice platform that flags anomalies
A non-developer used Replit and Claude to build an invoice platform that flags high-value invoices, sends payment reminders and catches duplicates; on 130 demo invoices it found four anomalies.
Build a report-writing program for real estate appraisals
A real estate appraiser used Claude to build a report-writing program that removes tedious work from the appraisal process.
Run a solo company with eight agent departments
A founder set up CEO, CFO, COO, lawyer, accountant, marketing, CTO, and improver agents that share memory and consult each other.
Grade papers and build lessons with Claude Cowork
A teacher says Claude Cowork saved hours on curriculum design, grading papers, making PowerPoints, and analyzing student grades.
Draft follow-up emails automatically when no reply arrives
A user set up a Lindy flow that watches sent emails, waits for a reply, and drafts a follow-up if none comes.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.