Bijblijver

Products

What we build for ourselves.

Products and tools made at Bijblijver. They are how ideas get tested before they reach a client, and some of them run underneath the websites we ship.

00 / Products 51°55'N 4°28'E Rotterdam
P

Products and engines

01 / 04   SaaS, pre-launch

Sparton

Know what your competitors changed this week.

Sparton finds the shops you compete with, reads their whole catalogue from their Shopify, WooCommerce and Lightspeed feeds, matches it to your products and writes a short report every Sunday. Every number links to the page it came from.

For
NL and BE webshops
Built with
Python, FastAPI, Postgres
Tests
581
Source on GitHub ↗
sparton, landing page01 / 08
The Sparton landing page, scrolled from top to bottom

02 / 04   Engine, under Sparton

shopfeed

Exact prices from the shop's own feed, no scraping.

Most small webshops publish their whole catalogue as JSON. shopfeed reads it from Shopify, WooCommerce and Lightspeed, snapshots it and lists what changed: prices, sales, stock, new and removed products. No LLM calls. It is the engine under Sparton.

Reads
Shopify, WooCommerce, Lightspeed
Built with
Python
Runs under
Sparton
shopfeed, example data03 / 03
$ shopfeed detect competitor.nl
shopify
$ shopfeed snapshot competitor.nl monday.json
shopify: 212 products -> monday.json
$ shopfeed snapshot competitor.nl tuesday.json
shopify: 212 products -> tuesday.json
$ shopfeed diff monday.json tuesday.json
price    Candle: 24.95 -> 19.95 (-20.0%)
sale     Candle: on sale (was 24.95)
stock    Mug: sold out
new      Vase: new product at 34.50 EUR
removed  Lamp: no longer listed
5 change(s)

03 / 04   Lab

Iris

A voice agent that runs this machine.

Iris listens for a wake word, reads the screen through Windows UI Automation and asks before it changes anything. The risk tiers live in code, not in the prompt. Fully local: no cloud, no cost per command.

Built with
Local LLM, UI Automation
Tests
150
Status
Running in the lab
iris, dashboard01 / 01
The Iris dashboard

04 / 04   Engine, under Vistora

Fidelity

Did the AI change the room, or only the light?

Models that relight property photos sometimes invent things: a window on a blank wall, a longer kitchen. Fidelity aligns each output with its source and flags structural changes, so a person only reviews the photos that need it. On the screen: a staged test, the same loft relit with a window added to the brick wall.

Checks
AI-edited property photos
Built with
Python, OpenCV, NumPy
Speed
About 0.4 s a photo, CPU only
fidelity, staged test pair01 / 03
Fidelity checking a staged test pair: a loft photo and its relit copy with a window added to the brick wall, the structure Fidelity compares, and its verdict FAIL with the added window flagged
C

Contact

Let's build
something.

jaydelauw@gmail.com