Devin Huber
n8n · 14 nodes

House Search Daily & Weekly Digest

Turns noisy listing-alert emails from three sites into a star-rated daily digest, plus a Sunday email with the week's best ten and an interactive board.

For a businessTurns a flood of duplicate listing alerts into a ranked shortlist, so the buyer only looks at homes worth a visit. The same scoring pattern fits leads or inventory.

Built for
A family member's house search (personal)
Model
None, fully rule-based
Stores
Listing ledger (Data Tables)
Schedule
Daily 8am · Sunday 8am
126point weighted score
13scoring factors
3-passdeduplication

01 · ContextThe problem

Listing alerts from Zillow, Redfin and Realtor.com arrive as separate, noisy emails: the same house three times, auctions and foreclosures mixed in, and no way to compare homes against what the buyer actually cares about.

02 · PipelineHow it works

  1. Fetch alert mail

    Reads the day's alert emails from a dedicated inbox.

  2. Parse and filter

    Extracts listings from all three sites and blocks auctions and foreclosures with a 15-term filter that logs the phrase behind every drop.

  3. Deduplicate

    Three passes against the ledger: listing ID, then normalized address plus ZIP, then across sources in the same batch. Merged records keep the longer description.

  4. Score out of 126

    13 weighted factors, led by area and safety (32), value per dollar (26) and condition (20), plus lot, size, proximity to family, street setting, parks, HOA, commute, beds, baths and walkability.

  5. Daily digest

    A star-rated email goes out and the ledger is updated. Five stars starts at 65.9% of the maximum.

  6. Sunday board

    The best 10 of the last 7 days, emailed with an interactive board attached.

03 · ChoicesDesign decisions

No LLM on purpose

Scoring homes needs to be cheap, repeatable and explainable, so every point comes from a rule you can read.

Reading the listing text

Weighted condition phrases like "new roof" (+3.0), "updated kitchen" (+2.5), "as-is" (−6.0) and "investor special" (−5.0) scale to a 0–12 updates score. Distances use haversine math.

Missing data is flagged, not punished

Missing fields score at the midpoint and are flagged on the card.

Suspiciously cheap gets a warning

Anything under $112 per square foot gets a "verify not an auction" chip.

04 · EvidenceHow I tested it

Runs daily against real alert mail, with every auction drop logged so the filter can be checked.

05 · Honest notesKnown limits

  • Alert emails don't include HOA fees, so the HOA factor is fixed.
  • It relies on alert emails rather than scraping, so coverage depends on the saved searches.

Want to see this one run?

I'll screenshare the workflow, the test set and the results, including what didn't work.

Get in touch