Build a nightly real-estate deal scout

A builder used Claude Code to scan MLS listings overnight, underwrite the ones that fit, and email shortlisted deals before morning.

Done with Claude Code· Advanced· Jul 2026

What they did
They set up a Claude Code agent on a nightly Inngest schedule. It pulled MLS data from a property-data API, filled in missing beds, baths, and square footage from a second endpoint, then screened listings by price and cap rate before running the remaining deals through several AI models in parallel. The system also watched user signals like saves, passes, and closes to improve its fit scoring over time.
How it went
The project turned into Dealsletter, which reached almost 200 users and more than 2,400 newsletter subscribers. The hardest problems were incomplete listing fields, learning from too little early feedback, and duplicate properties showing up in multiple feeds, so they had to add confidence checks and reconciliation logic.
Tips
If your data comes from multiple sources, plan for missing fields and duplicate records from the start. Before any email goes out, have the agent show a dry run of the deal list and wait for approval.

Try it yourself with Claude Code

Create a nightly deal-scouting agent for [market]. Each night, pull new MLS listings from [source or feed], filter them by [price], [property type], and [cap rate], then underwrite the ones that pass using [your rent, expense, and financing assumptions]. If key fields such as beds, baths, or square footage are missing, try to fill them from [secondary source]. Score each deal, reconcile duplicates across feeds, and show me a dry run of the shortlist first. Pause and ask for my approval before sending any email to [recipient] at [time].

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on reddit.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.