Frequently asked questions

Product

How fresh is the data?

Each uncached lookup fetches the live listing at the moment you ask, taking around four seconds. Responses are then cached for five minutes, so a repeat lookup of the same item can be up to five minutes old. For fast-moving items, treat every response as a point in time rather than a settled fact.

Why did the same item return a different price an hour later?

On marketplaces with third-party sellers, the buy box rotates between sellers who set their own prices and hold their own stock. A price of $12 from a seller with nothing in stock and $19 from one who can ship today are both correct answers at different moments. Use seller.name and availability.transactable_offers to tell them apart.

Do you provide search, or only individual products?

Individual products today, by item ID or URL. Search and category browsing are not available yet.

How reliable is it, honestly?

Roughly 1 in 20 uncached lookups fails and returns a 502. You are not charged for those, and a single retry almost always succeeds. Build one retry into your client and the practical success rate is very high. We publish this rather than bury it because you will meet it in production.

Can I get bulk exports or a full catalogue dump?

No. Every lookup is a live fetch, so there is no warehouse to export from. This suits monitoring a known set of items, not harvesting a whole category.

Billing

What exactly costs a credit?

Any lookup that answers your question, including “no product exists at that identifier”. Failures on our side — blocked fetches, timeouts, upstream errors — are free and refunded automatically. Check X-Credits-Charged on any response.

Cached responses still cost a credit?

Yes. You are paying for the answer, not for our infrastructure. The cache makes your integration faster; it is not a discount mechanism.

Do credits expire?

No. Buy once and use them whenever.

Can I get a refund?

Unused credits are refundable within 14 days of purchase. Consumed credits are not.

Coverage

Why are only a few marketplaces available?

Because only a few are finished. Each marketplace needs its own extraction and its own answer to that site's bot protection, and shipping a half-working one would cost you more in bad data than it saved you in waiting. The rest are listed with honest status so you can see what is coming.

Do all marketplaces return the same fields?

The envelope is the same, so code written against one marketplace works against another. The contents follow the marketplace: eBay listings carry auction state and seller feedback percentages, Walmart products carry buy-box sellers and star ratings. Anything a listing does not have comes back as null rather than being omitted. Each marketplace's page lists exactly what its lookups return.

When will Amazon be ready?

It is in development. We do not publish dates we cannot commit to. Join the waitlist on the relevant page and you will hear when it ships.

Can I request a marketplace?

Yes, and it matters — roadmap order is set by demand. Use the request form on any catalogue entry.

Technical

Is there an SDK?

No, and you do not need one. It is a GET request with a bearer token returning JSON. The docs have working samples in Python, Node, PHP and shell.

What are the rate limits?

3 requests per minute on the free trial, then 8, 20, 120 depending on the largest bundle purchased. Limits reflect real capacity — every request is a live page fetch, so throughput is the scarce resource.

Can I use this for high-volume catalogue work?

Not today. If you need sustained volume beyond the Scale tier, talk to us before building against it rather than after.

Is my API key recoverable?

No. Keys are stored as hashes and shown once. Lost keys should be revoked and replaced, which takes effect immediately.