hendeaux
Selected work

01 — Real estate · TypeScript · Railway

King MLS
listing automation

King MLS — Mike’s formula, put to work for his clients

Client
King MLS
Status
In production, used by the client
Access
Private to the client
Dashboard

1 / 2Dashboard

Open original image (opens in a new tab)
King MLS dashboard showing synced listings, price reductions, and saved searches

What was built

Mike King has an established formula for evaluating real-estate opportunities in Los Cabos. I built a system that applies his formula to saved MLS searches and brings the ranked listings and market information together on his website. He shares that information through his newsletter and with clients comparing properties and deciding what to buy. The supporting automation keeps listing data and ranked Google Sheets results up to date.

What the system includes

  • Applies Mike’s own ranking formula to his saved MLS searches
  • Brings property and market information together for newsletter readers and clients making decisions
  • Syncs listings from the Spark RESO API and sends ranked results to Google Sheets
  • Includes data checks, duplicate repair and alerts
  • Runs in production on Railway, with maintenance scripts that preview changes before applying them

From criteria to review

A shortlist with reasons behind it.

A local demonstration of the personalized-ranking candidate, using the actual interface and ranking code with a fictional buyer and four synthetic properties. It shows how Mike can inspect a match before deciding what to share.

01

Start with the buyer.

A $350,000 budget, at least two bedrooms and a pool. Ocean view is a preference. The profile separates what a property must satisfy from what makes it more appealing.

Buyer criteria · actual local interface with a fictional buyer and synthetic properties. September 10, 2026.

1 / 3Buyer criteria · actual local interface with a fictional buyer and synthetic properties. September 10, 2026.

Open original image (opens in a new tab)
Actual King MLS buyer editor with a fictional Rivera family profile: $350,000 budget, two bedrooms, required pool access and preferred ocean view

02

See why a property fits.

Mike’s market ranking stays intact. A separate match rank applies the buyer’s criteria: in this example, the third-ranked market listing becomes their first match. Each score shows its reasons.

Computed matches · actual local interface with a fictional buyer and synthetic properties. Market rank and buyer match rank remain distinct.

2 / 3Computed matches · actual local interface with a fictional buyer and synthetic properties. Market rank and buyer match rank remain distinct.

Open original image (opens in a new tab)
Actual personalized ranking table: fictional Palm Court ranks first for the buyer while retaining market rank three; Garden Place has an unverified pool warning

03

Keep the questions visible.

A property with no pool is excluded. One with missing pool information stays visible with a warning. Those are different decisions, and the interface keeps them available for Mike’s review.

Reasons to review · actual local interface with a fictional buyer and synthetic properties. No shortlist was approved or sent.

3 / 3Reasons to review · actual local interface with a fictional buyer and synthetic properties. No shortlist was approved or sent.

Open original image (opens in a new tab)
Actual King MLS review state showing three of four fictional properties eligible, one excluded for lacking the required pool and an explicit pending-confirmation policy

Separate pilot workflow

Prepare a buyer-facing draft.

A separate draft-preparation workflow turns buyer intake and a property snapshot into an email, a comparison page and a review brief. These are real generated outputs with fictional data. This pilot uses its own scoring; it is not an automatic handoff from the matching screen above. Mike still reviews the choices before sharing.

Email draft · generated locally by the existing pilot script from fictional intake and properties. No message was sent.

1 / 2Email draft · generated locally by the existing pilot script from fictional intake and properties. No message was sent.

Open original image (opens in a new tab)
Actual generated email draft for a fictional Rivera buyer, showing three invented properties with prices and fit notes
Buyer page · the separate pilot uses its own scoring, not the required-feature matching above. All properties and feed dates shown are synthetic; the packet still needs review.

2 / 2Buyer page · the separate pilot uses its own scoring, not the required-feature matching above. All properties and feed dates shown are synthetic; the packet still needs review.

Open original image (opens in a new tab)
Actual generated buyer shortlist with fictional property details, fit notes and pricing signals

Something like this for your business?

Start with the free plan, or a call.

Tell me which tasks take up your week and I’ll send a plan within 48 hours. Or book twenty minutes and we’ll talk through what a system like this would need for you.

Get a free automation planBook a discovery call
Previous projectFutrsport
storefront themeNext projectPost Grad Project
a connected business