I keep a short watchlist: a laptop I want under $700, a SSD under $60, a specific coffee that shouldn't cost more than $9. Checking five pages manually every morning is exactly the kind of boring work a small script does better.
Here's the whole watcher. Standard library only - no pip installs, no API keys, no paid data feeds.
import urllib.request, json, re, time
WATCHLIST = [
{"url": "https://example.com/product", "max": 700.00, "pattern": r"\$\s?([0-9,.]+)"},
]
def get_price(url, pattern):
req = urllib.request.Request(url, headers={"User-Agent": "Mozilla/5.0"})
html = urllib.request.urlopen(req, timeout=20).read().decode("utf-8", "ignore")
m = re.search(pattern, html)
return float(m.group(1).replace(",", "")) if m else None
state_file = "prices.json"
try:
state = json.load(open(state_file))
except FileNotFoundError:
state = {}
for item in WATCHLIST:
price = get_price(item["url"], item["pattern"])
if price is None:
continue
old = state.get(item["url"])
state[item["url"]] = price
if old is not None and price <= item["max"] and (old is None or old > item["max"]):
print(f"DROPPED: {item['url']} now {price} (target {item['max']})")
elif old is not None and price != old:
print(f"changed: {item['url']} {old} -> {price}")
json.dump(state, open(state_file, "w"))
Run it from cron (0 9 * * * python3 watcher.py) and it prints a line whenever a price crosses your target or just moves. Swap the print for smtplib if you want an email instead.
Three things I learned running this for a while:
-
The regex is the product. Every site hides its price differently - some in JSON-LD, some in a
<meta itemprop>, some in a span with a class that changes every deploy. Grab the page, search for the price pattern, and check it still matches after every run. - One page per site is enough. You don't need a crawler. You need the same page a human would open.
- Keep the state file small. URL -> last price. That's it. When the file is 10 lines you will actually maintain it.
Update: the watcher is now installable from PyPI - pip install price-monitor-kit gives you the hardened version as a price-monitor command (config file, CSV history, Telegram/email alerts, JSON-LD detection). Same idea as the script below, plus the flaky-network armor.
I packaged my hardened version (retry on flaky networks, per-site config, JSON-LD detection, a cron template, and a troubleshooting checklist) as Price Monitor Kit - it's one dollar, because I'd rather collect honest first feedback than set a high price: datafixtools.sell.app/product/price-monitor-kit
If you'd rather learn than pay, everything above is enough to build your own - the kit only saves you the debugging evenings.
I maintain the product linked above. Prices on real sites vary; test the regex against your target page before trusting an alert.
Top comments (0)