Real estate matching engine

Every listing, scored on what the buyer actually wants.

Scorimmo checks a listing against a buyer's criteria — the numbers, the location, the description and the photos — and returns a 0–100 score, justified criterion by criterion. Built for MLSs, portals and brokerage networks, through an API or our interface.

Illustrative example

Condo, 2 bedrooms

Listing scored for “the Tremblay family”

78

/100 · on the criteria evaluated

  • Price above $600,000Listing

    list price: $549,000

    not triggered
  • Metro within a 10-min walkLocation

    nearest station 6 min away

    8/8
  • Over 900 sq ftListing

    listing: 842 sq ft

    0/5
  • Renovated kitchenAI

    recent counters and cabinets, photos 3 and 5

    10/10
  • Balcony or terraceAI

    neither the description nor the photos mention one

    not evaluated
18 points earned out of 23 evaluated = 78. The unevaluated criterion is left out.

The problem

Buyers don't think in filters.

MLS search tools work with filters: every field is true or false, and a listing is either in or out. But “close to the metro”, “bright”, “not on the ground floor” or “a kitchen we won't have to redo” aren't in any field — and not every wish carries the same weight. Searches end up too broad — hundreds of results to sort by hand — or too narrow, missing the right listing over 20 sq ft.

Result

Filter search: In the list or not

Scorimmo: A 0–100 score with every criterion explained

What is read

Filter search: Listing fields

Scorimmo: Fields, location, description and photos

Priorities

Filter search: Every filter weighs the same

Scorimmo: Points per criterion, with deal-breakers kept apart

Missing information

Filter search: Listing dropped, or kept silently

Scorimmo: Criterion flagged “not evaluated”, left out of the score

The score

How a listing is scored

Each buyer has a criteria set. It holds two kinds of criteria, handled differently.

Exclusions

The deal-breakers. If any one of them is true, the listing is eliminated and scores 0, whatever else it has going for it.

E.g. price over budget, ground floor, no parking.

Scored criteria

The wishes, each worth a number of points according to its importance. A criterion that is met earns its points; one that isn't earns none.

Score = points earned ÷ points of evaluated criteria × 100

Three ways to evaluate a criterion

Each criterion states how it is checked. The engine runs the methods from least to most expensive.

  1. 1 · Instant

    Listing data

    Exact comparison on the listing's fields.

    • price, living area, rooms, bedrooms, bathrooms
    • floor, year built, lot size
    • condo/HOA fees, property taxes, energy rating
  2. 2 · Fast

    Location measurements

    Computed from the property's coordinates, using Google Maps Platform, our own copy of OpenStreetMap and public open data.

    • walking minutes to the metro, a grocery store, a school
    • commute to an address by car, transit, bike or on foot
    • distance to the sea or a park, density of shops
    • air quality, pollen, rooftop sunshine
  3. 3 · In-depth

    AI reading

    For what no field says. The model reads the description and, when the criterion calls for it, the photos, then states what it relied on.

    • renovated kitchen
    • bright rooms
    • open view
    • no carpet

Criteria

How criteria are created

Criteria come from the buyer, in their own words. Three ways to get them, which can be combined:

  1. A free-form brief

    The broker pastes their notes, an email or a call summary. The AI proposes a list of criteria — exclusions, points, evaluation method — each with a sentence saying how it will be checked. Nothing is saved until it has been reviewed.

  2. A questionnaire for the buyer

    A personal link sent to the buyer: budget, size, rooms, floor, nearby places, commutes. Their answers become proposed criteria, reviewed and applied in one click.

  3. Direct editing

    Add, change, weigh or remove a criterion, in the interface or through the API. Start from a market-specific template.

  4. Amounts and areas use the market's currency and unit: dollars and square feet in Quebec, euros and square metres in France.
Example

Broker's notes

“$650,000 max. Not on the ground floor. Walking distance to the metro matters a lot. Three bedrooms, and ideally a yard for the dog.”
Proposed criteria
  • ExclusionPrice above $650,000Listing
  • ExclusionGround floorListing
  • 120 ptsMetro within a 10-min walkLocation
  • 80 pts3 bedrooms or moreListing
  • 40 ptsPrivate yard or lotAI

Who it's for

The different use cases

Scorimmo doesn't publish listings. It plugs into the platforms that already distribute them, or is used directly, by brokerage teams and individuals alike.

MLSs and real estate boards

Give your members match-based search on your own listings, next to your existing tools, without replacing anything.

Portals and software vendors

Add the score to your product: your screens and your users, our engine underneath.

Brokerage networks and brokers

Use the Scorimmo interface as is to qualify listings for each buyer client.

Individuals

Use Scorimmo to search faster and more effectively on for-sale-by-owner listing sites.

Integration

Two ways to work with us

The same engine, whether you build it into your platform or use ours.

Through the API

You keep the display, we do the scoring.

  • One token per company, revocable at any time; we only keep a fingerprint of it.
  • Contract quotas enforced: analyses per month, criteria sets, requests per minute.
  • A geographic perimeter: the areas your contract covers.
  • Listings in RESO format, or in your native format through an adapter we set up with you.
  • One link per record: the buyer answers the questionnaire or describes their search (text, PDF, voice note), and their criteria come back to you by webhook.
  • Up to 100 listings per search, scores and details returned as they come or by webhook.
  • Free re-analysis for 24 hours when neither the listing nor the criteria changed.
POST /v1/partner/analyses/batch
Authorization: Bearer smk_live_…

{ "criteria_id": "sc-3f9a…",
  "properties": [ { … }, { … } ] }

→ { "ranking": [ { "score": 78, "result": { "breakdown": [ … ] } } ] }
Read the API documentation

In the Scorimmo interface

For teams who want to use it as is.

  • One file per buyer, with their exclusions and scored criteria.
  • Criteria drafted from notes, or from a questionnaire sent to the buyer.
  • Listing analysis with every criterion detailed and the sources used.
  • Access on approval, with an analysis allowance set per account.

Getting started

  1. 01

    A first conversation

    Your market, your volumes, your listing format and what you want to display.

  2. 02

    Setup

    Market, perimeter, quotas and token: all configured on our side.

  3. 03

    Integration

    You call the API from your platform, or your teams use the interface.

Principles

What we don't compromise on

No guessing

Missing information means “not evaluated”, never a made-up answer.

Controlled costs

Cheapest methods first, stop at the first exclusion, results reused for 24 hours.

Hosted map data

A copy of OpenStreetMap per market, on our servers, refreshed automatically every month.

AI-vendor independent

The language model is interchangeable; no criterion depends on a specific one.

Isolated data

Each client only sees its own criteria sets and its own analyses.

Multiple markets

Currency, units, questionnaire and data sources adapted to each market, in Quebec and in France.

Let's talk about your platform

One meeting is enough to see whether Scorimmo fits your listings and your tools.