Watch Apple Refurbished store pages and get notified the moment a machine you care about appears — with a small web dashboard showing what's currently available plus the full appearance / price-change history.
Zero-setup: runs on SQLite out of the box — no database to install.
- Scrapes any Apple Refurbished page(s), any locale — Mac Studio, Mac mini, MacBook Pro… whatever URLs you configure.
- Parses Apple's embedded
application/ld+json(with an HTML fallback): part number, title, price, RAM, storage, chip, CPU/GPU cores. - Diffs against the previous run and records every appeared, disappeared and price_changed event.
- Notifies via ntfy — but only for products matching your filters (keywords, min RAM per category, optional price ceiling). Notifications are entirely optional.
- Dashboard (Flask): what's available now + recent events + scraper health.
- SQLite by default, MariaDB/MySQL optional.
git clone https://github.com/lolfr/refurb_watcher.git
cd refurb_watcher
./scripts/install.sh # venv + deps + config.toml
$EDITOR config.toml # choose the URLs + filters you want
./venv/bin/python watcher.py # first run (creates refurb.db)
./venv/bin/python web.py # dashboard on http://127.0.0.1:5060Then schedule watcher.py to run periodically — see
scripts/crontab.example.
Everything lives in config.toml (copied from config.toml.example):
| Section | What |
|---|---|
[apple] urls |
The refurb pages to watch (any Apple locale). |
[database] backend |
"sqlite" (default) or "mariadb". |
[ntfy] |
Push notifications — leave url empty to disable. |
[filters] |
Which new products trigger a notification (keywords + per-category rules). |
[dashboard] |
Flask host/port. |
Set [ntfy] url to an ntfy topic (e.g. https://ntfy.sh/your-topic) and install
the ntfy app on your phone. Add a token only if your topic requires auth. Leave
url empty and the scraper just records to the database silently.
[database]
backend = "mariadb"
host = "localhost"
user = "refurb"
password = "…"
database = "refurb_watcher"Then pip install pymysql and apply schema.sql once. The SQLite
backend creates its tables automatically, so schema.sql is only for MariaDB.
watcher.py fetches each configured URL, parser.py extracts products
(primarily from the ld+json Product blocks Apple embeds, with an HTML
fallback), db.py diffs against the stored snapshot and records events, and
matching new products are pushed via ntfy. web.py serves a read-only dashboard
over the same database.
Apple changes its markup without notice, so the parser is deliberately defensive (an explicit error beats silently-wrong data), and the watcher never purges the snapshot when a fetch returns zero products (anti-scrape / transient errors).
AGPL-3.0 — © 2026 lolfr.