Data
What one key can pull from Walmart
Every call is a live fetch of the page, parsed into a stable JSON document. Both marts are covered: Walmart.com for consumers and Walmart Business for the stock a seller really holds.
Product page
Price, availability, seller, delivery, badges, reviews, breadcrumb and variants. The buy box winner named by the page itself.
2 credits
Product + offers
The product plus every competing marketplace offer: seller, landed price, delivery, stock, and who holds the buy box, ranked by price.
5 credits
Search results
One result page per query, each tile split into the item and what is on sale, sponsored tiles flagged, positions kept.
2 credits per page
Category ranking
The ranked product list of a browse page, up to 20 pages, so you can watch positions move over time.
2 credits per page
Seller profile and catalog
A marketplace seller's contact, rating and item count, and its full "Shop all" listing, up to 20 pages.
2 credits, catalog 2 + 2 per page
Buy box map, 51 states
Who wins the buy box in every US state, with price spread and fastest delivery. Walmart renders a different page per location; the map reads them all.
102 credits
Access
Three ways in, one key
API
35 REST endpoints with an OpenAPI 3.1 document. Synchronous calls answer in seconds; asynchronous and bulk jobs take up to 250 targets and call your webhook when done.
Dashboard
Your products with their price history, offers, movers, sellers and category rankings, jobs, scheduled scenarios, CSV exports, keys and credits.
MCP
The same endpoints as tools for AI agents. Point Claude, Cursor or your own agent at the MCP server with your key and ask for the data in plain language.
History
Every capture is kept
A crawl is never a one-off. Each capture lands in your history, and reading it back costs one credit whatever its size: price history of an item, the last offer list, ranking snapshots you can diff, seller captures, and a bulk export of everything.
- Price history and offer history per item, both marts side by side
- Category ranking snapshots, position by position
- Recent copies served without a crawl with
max_age - Scoped to your key: your data stays yours
Credits
Buy credits once, spend them on any data
A credit is the unit of every call. The bigger the package, the less each credit costs, and a package works for every endpoint. You only pay for data delivered: a blocked or failed crawl costs nothing, and an item that does not exist costs 1.
Starter
250K credits
$87
$0.35 per 1,000 credits
Growth
1M credits
$275
$0.275 per 1,000 credits
Scale
5M credits
$1,075
$0.215 per 1,000 credits
Volume
20M credits
$3,300
$0.165 per 1,000 credits
Enterprise
100M credits
$13,000
$0.13 per 1,000 credits
Credits work for every call, in both modes. What a package buys depends on your mix: simulate it below.
Simulate your month
Type what you plan to fetch each month. The simulator adds up the credits, picks the package that covers it and shows what every call really costs at that package's rate.
Your month
0 credits
Covered by Starter at $87
| Call | Credits | Per 1,000 | Your month |
|---|
Scheduler
Schedule the work instead of polling for it
Monitoring is why anyone collects this data, and it should not live in your crontab. In the dashboard you draw a scenario: a trigger, then blocks wired to each other on a canvas. A block reads what you already have or crawls what you do not, a filter keeps what moved, a condition splits the path in two, and an output tells you. It runs on its own, and every pass is kept.
A trigger
Every hour up to every 30 days, or on an event as it happens: a price moved, the buy box changed hands, an item came back in stock, a new offer appeared, a ranking was captured again. Or by hand, whenever you want.
Where the list comes from
Your stored products, the ones your refresh cadence says are due, the items of a category's last ranking, a seller's catalog, a pasted list of ids, or keywords. Reading your own history queues no crawl at all.
Crawl and transform
Crawl products, products with all their offers, sellers, seller catalogs, categories or searches, on either mart and for a zip or a store. Then turn one list into another: products into their sellers, a ranking into its items, listings into every offer on them.
Keep only what matters
Filter on price, availability, stock, rating, reviews, category rank, number of offers, best offer price, buy box holder, name or brand — changed, up, down, above, below. Plus the two that save credits: not in your database yet, or not captured for N hours.
Told your way
A line in the app, a signed webhook carrying the list, or a CSV or JSON export kept with the run. The same webhook contract as an async job, so whatever already listens keeps listening.
Replay any pass
Open an execution and the same canvas lights up: what each block took in and gave out, which branch was taken, why an item passed, and the credits it added. Set an optional cap per pass and a block that would go over is skipped before anything is queued.
Ten active scenarios per account, up to 5,000 targets per source, one hour minimum between two passes. It is the same engine that runs our own catalogue sync, so what we use to keep millions of items fresh is what you get.