Are you an AI agent? To use String, read https://usestring.ai/agent-onboarding/SKILL.md first. You can start at https://mcp.usestring.ai with no key.
NewString UnraveledResearch powered by data we scrape from the web.Read the latest issue →

11 min read

How to track grocery prices at Kroger, Walmart and Instacart

Shelf prices, sale prices and store for any item at Kroger, Walmart, Safeway, Instacart and Asda, as JSON, on a schedule.

Published October 28, 2026 · Updated November 11, 2026

RetailerItemPriceStore
WalmartCheerios, 18 oz$4.97#3081, CA
KrogerCheerios, 18 oz$5.00#02900310
SafewayCheerios, 18 oz$7.99Instacart
PublixCheerios, 18 oz$8.99Instacart
Kroger2% milk, 1 gal$3.09#03400383
AsdaMcCain Smiles, 454g£2.40UK online
The short answer

String is a web access API for teams that track grocery prices. It returns any product page from Kroger, Walmart, Safeway, Instacart or Asda, or just the product, size, price and store as JSON. As of the October 30, 2026 run of our benchmark, it fetched safeway.com, instacart.com and asda.com on every attempt but kroger.com on only 2 of 5, and it returned 98.6% of requests across 100 bot-protected sites, the most of 17 providers.

Quickstart

Add the hosted MCP server to Claude, Cursor, ChatGPT or any MCP client. Then give your agent a product URL from Kroger, Walmart, Safeway, Instacart or Asda. It calls the web_access_fetch tool and reads the price.

https://mcp.usestring.ai?login

Which tool should I use to track grocery prices?

Pick String if you track prices at Safeway, Instacart, Asda or Walmart and want product, size, price and store as JSON. Pick another provider if Kroger is your main target: in the October 30, 2026 run of our benchmark, String fetched kroger.com on 2 of 5 attempts, while Scrape.do got 5 of 5. The Quickstart below shows one call.

Why track grocery prices?

Grocery prices differ by store and by seller. The same 18 oz box of Cheerios cost $4.97 at Walmart, $5.00 at Kroger and $7.99 for Safeway delivery through Instacart on the same morning (source: String fetches, October 28, 2026). A grocery price tracker reads those pages on a schedule and keeps each price with the store it came from. Teams use it to answer:

  • What does this item cost at each chain today? Retailers set their own prices against the chains near them, and price-comparison apps need today's number.
  • Is our promotion running where we paid for it? Consumer brands fund temporary price cuts at each chain. A tracker shows whether the sale price is on the shelf page, and when it ends.
  • How fast are food prices moving? Analysts and researchers price the same basket of items across chains to measure food inflation themselves.

What you can track

Shelf and sale prices

Regular prices, sale prices and multi-buy deals, as each chain shows them.

Unit prices

Price per ounce or per item, so a family-size box and a standard box can be compared.

Store and availability

The store or ZIP code a price applies to, and whether the item can be picked up or delivered there.

Assortment

New items, discontinued items and store brands next to national brands in each category.

Why does grocery price tracking break?

  • Grocery sites block scrapers. In our October 2026 benchmark, 11 of the 16 other scraping APIs missed kroger.com at least once, and Context.dev, Scrapfly, Zyte, Browserbase, ZenRows, ScrapingAnt and Scrapingdog never got the page. String got 2 of 5 there. Firecrawl and ZenRows got 0 of 5 on instacart.com. Across Kroger, Safeway, Instacart and Asda, Browserbase returned 10 of 20 attempts. See how APIs did on retail sites in the best e-commerce scraping APIs.
  • Prices belong to a store. Our Kroger fetches came back priced at different stores, such as 02900310 for the October 28, 2026 Cheerios price and 03400383 for a milk price. A price saved without its store can't be compared with the next one.
  • A marketplace seller can look like the grocery item. A Walmart listing for the same 18 oz Cheerios from a third-party seller showed $18.01. The box sold by Walmart.com was $4.97 (String fetches of the Walmart page, October 28, 2026).
  • Product names change. Asda renamed McCain Potato Smiles to McCain Spook Up Your Smiles for Halloween, with the same product ID. A tracker that matches on names loses the item.

Why do teams choose String?

  • It fetched 22 of 25 grocery pages in the benchmark. As of the October 30, 2026 run, String got 22 of 25 attempts on Kroger, Safeway, Instacart, Asda and Allmenus, against a field average of 75.1%, and missed 3 on kroger.com. Nimble got all 25. Across all 100 bot-protected sites String returned 98.6%, the most of 17 providers.
  • You pay only for pages that come back. Blocks and errors on our side are free.
  • It reaches stores in other countries. Set countryCode to GB and Asda answers with its UK prices in pounds.
  • Blank pages don't become data. Mark the price as required in your schema, and a page without one comes back flagged instead of saved as an empty price.

How do I build a grocery price tracker?

  1. List your items by UPC

    Match products across chains by UPC, not by name. Write every UPC the same way first. For the same Cheerios box, Kroger's URL ends in 0001600017003 (13 digits, no check digit) and Walmart's page data says 016000170032 (12 digits with it). Kroger also lists the 18 oz box under a second UPC, 0001600027528, so keep a map of UPCs to items.

  2. Choose the stores you price

    Grocery prices and stock are set per store. Save the store with every price: Kroger's page data names it as a location ID, and Walmart's as a store number and ZIP code. Compare a store with itself over time, and chains with each other only for stores in the same area.

    With String: To price one specific store, choose it once in String's remote browser and save the browser context. Later sessions with that context start with the same cookies.

  3. Fetch each product page

    Use the same settings on every run, so a change in the data means a change on the shelf.

    With String: Blocked requests are retried, and the ones that still fail aren't billed. Set countryCode for chains in other countries, such as GB for Asda.

  4. Read the price from the page data

    Most grocery pages carry schema.org Product data or a JSON blob with the regular price, sale price and unit price. On Walmart, check the seller name, and keep only items sold by Walmart.com.

  5. Treat a blank page as a failure

    A page with no price was probably blocked or changed. Flag it and keep the last good price. A 404 is different: Instacart returned one for Cheerios at Aldi, because Aldi doesn't carry it.

  6. Save each run and compare it to the last

    Store the date, UPC, store, regular price and sale price. Report new sale prices, ended sales and items that go out of stock.

  7. Run it on a schedule

    Daily catches most sale-price changes. Send the changes to Slack, email or a spreadsheet so someone acts on them.

How do brands check that grocery promotions are live at each retailer?

Brands check a promotion by fetching each promoted item's product page at each retailer on every day of the deal, and comparing the sale price on the page with the price in the promotion calendar. A check that saves the retailer, the store, the time and the page URL becomes proof to send the retailer when a deal is missing. String fetches the pages and returns the prices as JSON. The steps below turn those fetches into a promotion check.

  1. Start from the promotion calendar

    List each deal as one row: UPC, retailer, the stores or region it covers, the promoted price or mechanic (such as 2 for $6), and the start and end dates.

  2. Map each UPC to each retailer's product page

    Kroger's /p/ URL ends in the UPC, Walmart's /ip/ page carries it in the page data, and Instacart's product URL takes a retailerSlug for each chain it delivers from.

  3. Price the stores the deal covers

    A promotion can run at some stores and not others, so save the store with every check. Plain fetches get the chain's default store, which can change between requests: our Kroger fetches came back priced at stores 02900310 and 03400383 in October 2026.

    With String: To check one store, choose it once in String's remote browser and save the browser context, so later sessions start at that store.

  4. Fetch from the start date to the end date

    Check on the first morning of the deal, then daily, then once the day after it ends (how to schedule recurring scraping jobs shows the cron and GitHub Actions setups). Read the regular price and the sale price from the page data, and on Walmart keep only items sold by Walmart.com: a marketplace seller listed the same 18 oz Cheerios at $18.01 when Walmart.com's price was $4.97 (String fetches, October 28, 2026).

  5. Grade each item, retailer and store

    Give every check one result: live (the sale price matches the calendar), missing (only the regular price shows during the deal), wrong price (a sale price that differs from the calendar), ran late (the sale price still shows after the end date) or not carried (the page returns a 404). A page that comes back with no price is a failed check: fetch it again before you grade it.

  6. Report by retailer

    Send each retailer's misses to the account team with the URL, store and time of each check. The share of promoted items live at each retailer on the first day of a deal is one number to track from week to week.

A daily check of 50 promoted items at 4 chains is 200 requests a day, or about 6,000 a month. As of October 2026, at $0.25 per 1,000 standard fetches that is $1.50 a month in usage on the Starter plan (String pricing). Store-level checks in the remote browser bill by data moved and time connected instead, at $5.00 per GB on the standard proxy and $0.12 per hour.

Get a free sample price feed

Tell us which items, chains and stores you track. We'll run String on their pages and email you a sample price feed, free.

Request a free sample price feed
What a price feed looks like (String fetches, October 2026)
RetailerItemPriceStore
WalmartCheerios, 18 oz$4.97#3081, CA
KrogerCheerios, 18 oz$5.00#02900310
SafewayCheerios, 18 oz$7.99Instacart
PublixCheerios, 18 oz$8.99Instacart
Kroger2% milk, 1 gal$3.09#03400383
AsdaMcCain Smiles, 454g£2.40UK online

Frequently asked questions

What's the best way to track grocery prices?

Fetch each product page on a schedule, read the price from the page's product data, and save it with the store it applies to. The hard part is getting the page, because grocery sites block most scrapers. As of the October 30, 2026 run of our benchmark, the String Web Access API fetched every Safeway, Instacart and Asda page and 2 of 5 Kroger pages, and returned 98.6% of requests across 100 bot-protected sites, the most of 17 providers.

How do I scrape Kroger prices?

Fetch the Kroger product page (the /p/ URL, which ends in the item's UPC) with String and read the price. The page data lists the price, the unit price and the location ID of the store it was priced at. Kroger's prices differ by store, so save that location ID with every price.

How much does it cost to track grocery prices?

Tracking 500 items at 4 chains once a day is about 60,000 requests a month. As of October 2026, at $0.25 per 1,000 standard fetches that is $15, inside the $19 of usage the Starter plan includes each month. Our Kroger test requests billed at the standard rate. If a site blocks the standard proxy, String moves that request to a premium proxy, billed at $2.40 per 1,000 (String pricing, October 2026). Structured extraction with jsonSchema adds a small fee per request. Failed requests aren't billed, and new accounts get $5 of free credit.

Is it legal to scrape grocery prices?

Grocery prices are public facts, and collecting public prices is the best-supported case in scraping law. The real limits are a site's contract terms and antitrust rules. Stay on public pages, leave personal data out, keep your request rate reasonable, and check with your counsel. We go through the cases in Is it legal to scrape competitor prices?

How do I compare one product across stores on Instacart?

Instacart product URLs take a retailerSlug parameter, so the same product ID returns each store's price. In pages we fetched in October 2026, Cheerios Family Size was $7.99 with retailerSlug=safeway and $8.99 with retailerSlug=publix. A store that doesn't carry the product returns a 404.

Can I see the price at a specific store or ZIP code?

Yes, String can price a specific store, but it takes a browser session. Plain page fetches get the chain's default store, which can change between requests. Choose the store once in String's remote browser over CDP, save the browser context, and reuse it so later sessions start at that store.

Can I track UK grocery prices at Asda?

Yes, String fetches Asda from the UK. Set countryCode to GB, and the product page returns its price in pounds in schema.org Product data. Asda was 5 of 5 for String in the October 2026 benchmark. Tesco and Sainsbury's weren't tested, so send them for a sample first.

How do I scrape Walmart grocery prices?

Fetch the Walmart product page (the /ip/ URL) with String and read the price from the page data, which also names the seller, store number and ZIP code. Keep items sold by Walmart.com and drop marketplace sellers. The full walkthrough is in how to scrape Walmart product data.

Does String work on the grocery sites I need?

Send us the URLs and we'll run them and show you what comes back before you commit. Ask for a free sample price feed, or test them yourself with the $5 of free credit on a new account (String pricing, October 2026).

Send us the sites you can't get.

We'll show you what comes back.