Watch any product's price with OpenClaw

The advice everyone gives is to wait for the price to drop, and the reason almost nobody does is that watching is tedious. You add the thing to a cart, you check back a couple of times, you forget, and then you either pay full price out of impatience or buy it the week before a sale you never knew was coming. The tools that solve this, the price history sites and the browser extensions, mostly only cover Amazon, and the good ones charge for the parts you want.

Your Operator.io agent can do the watching for you, on whatever store the thing is sold. You send it a link and a price you would be happy to pay, and it checks once a day, keeps a record of how the price moves, and tells you the moment one of your items drops to your number or hits a low it has not seen before. The rest of the time it stays quiet.

How it works

You send the price watcher prompt, and the agent sets up a watch list and the daily check that runs it:

• When you send a link and a price you would pay, it reads the page, confirms the price it found, and adds a row to a CSV in its workspace with the item, link, store, your target, the latest price, the lowest it has seen, and the date it last checked.
• Every 24 hours it revisits each link, reads the current price, and writes it down, lowering the recorded low whenever something drops past it.
• It messages you only when an item reaches your target or sets a fresh low, so a week where everything sits flat is a week you never get pinged.

That recorded history is what makes the daily check mean something. Without it the agent would only know today's price, when the useful question is almost always how today compares to the last few weeks. You can see and pause the run any time from the Automations page in your dashboard.

Because the agent reads the price from the page itself rather than from one store's catalog, this works on any site that shows a price publicly. A link to a camera on a specialty retailer, a coat on a brand's own site, a monitor at a big box store, all of them work the same way, since all the agent needs is a page with a price on it. Sites that hide their prices behind a login or a cart it cannot fill are the exception, and when one of those trips it up, it tells you rather than logging nothing.

The prompt

This is the instruction the agent acts on:

Be my price watcher. Create a spreadsheet in your workspace, a CSV with columns
for the item, the product link, the store, my target price, the latest price,
the lowest price you have seen, and the date you last checked. When I send you a
link, check the file first so the same item never lands twice, then open the
page, read the current price, and add a row with the target I give you. Once a
day, visit each link, record the price, and lower the all time low when something
drops past it. Set up an automation that runs every 24 hours and messages me here
only when an item reaches my target or sets a new low, so quiet days stay quiet.
Before you start, ask me which products to watch and the target price for each,
and what time of day to run the check.

The same prompt is waiting in the prompts library, so you can send it to your agent without typing it out.

Using it day to day

Adding something to the watch list is as simple as pasting a link in your channel and saying what you would pay for it. The agent reads the page, confirms the price it found so you can catch a misread before it matters, and starts tracking from there. Drop an item you have stopped caring about and it removes the row. Change a target after you have watched a price float for a while and it updates it. None of this needs a settings screen; you do it all by talking to the agent.

The record it keeps is the part a simple alert cannot give you. Because the history lives in a file the agent owns, you can ask it to reason about the data instead of just reporting it.

Whether a price is low or just lower than yesterday, how much something has bounced around over the past month, whether the dip you are looking at is the real floor or a number it beat three weeks ago, these are questions it answers from the rows it has been keeping. That context is what turns "it is on sale" into "this is the lowest it has been since you started watching, buy it now."

Why prices move and what to watch for

Retail prices are rarely static. Dynamic pricing, where a retailer adjusts a price based on supply, demand, season, or competitor moves, is standard in airlines, hotels, and ride hailing apps, and it has spread to everyday retail. The Future of Privacy Forum's 2026 report on data driven pricing describes how retailers increasingly use automated systems to change prices in real time based on market conditions, and in some cases based on signals about individual shoppers. Two people can see different prices for the same item on the same day, which is one reason a personal price log matters more than a single snapshot.

Amazon shoppers have had dedicated tools for years. CamelCamelCamel, founded in 2008, tracks price history on Amazon marketplaces across the US, UK, Germany, and several other regions, and its browser extension The Camelizer embeds charts directly on product pages. How-To Geek's guide walks through reading those charts and using the "Good Deal" and "Best Price" markers CamelCamelCamel assigns based on historical lows. The service is completely free, funded by affiliate commissions and advertising.

Keepa goes deeper. Its browser extension, used by over four million people according to independent reviews, tracks more than three billion Amazon products and shows price history charts, sales rank data, and Buy Box statistics directly on product pages.

The free tier covers price history charts, price drop alerts, and tracking up to 5,000 products. The paid Data Access subscription costs €19 a month or €189 a year and unlocks sales rank history, the Product Finder tool, Buy Box analytics, and API access for sellers who need programmatic data. For pure Amazon deal hunting, Keepa and CamelCamelCamel remain the most data rich options.

The gap appears when the thing you want is sold somewhere else. A camera at B&H Photo, a jacket on a brand's direct to consumer site, a GPU at a regional electronics retailer, a flight on an airline's own booking page: none of these appear in Amazon price trackers. That is where reading the price from the page itself, the way your agent does, covers ground the Amazon tools cannot.

Handling tricky pages

Not every product page is straightforward. Some retailers show a "starting at" price that changes when you select size or color. Others load the price with JavaScript after the page shell renders. Some show member pricing only after login.

When you add an item, tell the agent which variant you mean: "watch the 27 inch model in black" or "track the sale price, not the list price." If the page requires selecting options before a price appears, ask the agent to load it in a browser, pick the options, and then record the number it finds.

Sales tax, shipping, and currency add another layer. The agent logs the price it reads on the page, which may be before tax and shipping. If your target price includes those costs, say so when you set the target: "alert me when the total with shipping drops below $400." For international stores, note the currency in your target so the agent does not compare dollars to euros.

Flash sales and lightning deals can move a price for a few hours and then revert. A daily check catches sustained drops but may miss a brief window. If you are watching something that goes on flash sale regularly, tell the agent to check twice a day during sale season, or set a lower target that accounts for the typical sale price rather than the list price.

A page that read cleanly the day you added it can still drift later. Retailers redesign, move the sale price into a different element, or restyle a page for a promotion, and a read that was landing on the right number can start landing on the wrong one. The daily message is the backstop: when an alert fires, the agent tells you the price it saw and the link it read, so a number that looks off is one you can open and have it re check before you buy. If a site changes enough that it keeps misreading, describe where the price sits on the page now or ask it to read the page in a browser, the same fix that handles the JavaScript pages above.

Why a spreadsheet beats another price app

The tools people reach for first are built around Amazon, and the one that reaches further has a catch.

CamelCamelCamel is completely free, and Keepa goes deeper with the charts and sales rank data resellers rely on, but both stop at Amazon, so the moment the thing you want is sold somewhere else they cannot help. Honey, the PayPal owned coupon finder, casts the wider net, though its price history is shallow and its affiliate practices came under fire in late 2024. Marketing Brew reported on class action lawsuits alleging Honey swapped creator affiliate links with its own at checkout, a practice described as cookie stuffing in the Wikipedia entry on the controversy. Honey modified its extension in 2025 after Chrome Web Store policy changes, but the litigation is ongoing.

The agent works differently. Its log is a CSV you own, it follows whichever store sells the thing rather than one catalog, and it costs nothing beyond the OpenClaw you already run, with no monthly charge for watching a handful of items. It will not embed a chart on a product page the way Keepa does, and for pure Amazon deal hunting the dedicated tools still show more raw data. What it does well is watch the few specific things you plan to buy, on whatever sites sell them, and ping you only when one hits your number.

When you are ready, open the prompts library and send the price watcher to your agent. It asks which products to watch and the price you would pay for each, then it goes quiet until one of them hits your number.