“At least A3” means at least A3.

Spec matching for WooCommerce stores with specialty catalogues. Buyers describe what they need in plain language. Your database decides what matches, not the AI model.

I need a multi function printer that can print colour, A3 and do A3 scanning

Looking forcategory: multifunction printerscolour: yesmax_print_size ≥ A3max_scan_size ≥ A3

Three printers meet every requirement. Two more print colour at A3, but their listings don’t say what size they scan, so they’re shown separately.

3 full matches

Product photo

Ostra C3520 A3 colour MFPPrint A3 · Scan A3 · Colour

$3,490In stock

Product photo

Keldon WX-A3 ProPrint SRA3 · Scan A3 · Colour

$4,150In stock

Product photo

Ostra C4030iPrint A3 · Scan A3 · Colour

$5,2802 left

2 unconfirmed

UnconfirmedHalvar MP-7300Scan size not listed

$2,980

UnconfirmedKeldon WX-3100Scan size not listed

$3,120

An example store with example products. Prices and stock come live from the store at the moment the buyer asks. The model never writes a price.

When nothing fully matches

Near misses say exactly what’s missing

Shown only when no product meets every requirement, so buyers can decide what they’ll compromise on.

Near miss: scans at A4 onlyHalvar MP-5200$2,450

Near miss: mono onlyOstra M3000 A3$1,890

Example store data. Labels name the unmet requirement in the buyer’s terms.

How it works

The model reads and explains. The database decides.

While indexing, the model reads each product and extracts its specs into agreed fields, using your store attributes where they exist.
When a buyer asks, the model turns their request into filters.
The plugin runs those filters against the database.
The model explains the results, using only the specs that came back.

// request, as filters
category: multifunction printers
colour: yes
max_print_size: at least A3
max_scan_size: at least A3

// size order used for comparisons
A5 < A4 < A3 < SRA3 < A2

// a missing value
max_scan_size: unknown → unconfirmed, never “no”

Results buyers can trust, including B2B buyers

Unknown isn’t no

A missing spec stays unknown. Those products appear separately as unconfirmed.

Real comparisons

A3 ranks larger than A4, so “at least A3” is a true test, not a keyword.

Live store data

Price and stock come from the store when the buyer asks, on real product cards.

Store rules apply

Hidden products, variations, member pricing, currencies and tax display are respected.

Specs you can check

Approve the fields and correct values. Your edits survive re-indexing.

Your own AI key

Same engine and usage limits as Search and Knowledge.