kwery is price data infrastructure for e-commerce. We deliver decision-ready marketplace pricing data via a native API, real-time change events, health signals, and full auditability. You stay in control of your repricing logic; kwery is the reliable data layer underneath it. How it works.
Repricing platforms, enterprise retailers, and multichannel commerce teams that need reliable, normalized marketplace price data without building and maintaining their own collection pipelines. Who it's for.
Sign up free in seconds, every new account gets a one-time grant of 1.000 credits and no credit card is required. Upgrade to a Growth plan when you need more volume. Enterprise teams can request a scoped contract with dedicated support and custom sources. Get API Key.
Generate an API key, call POST /stream with a GTIN list and a callback_url, and results arrive as JSON via webhook. Most teams are live within a few days using copy-paste examples in cURL, Node.js, and Python. API documentation.
Three things: an account, an API key generated in your developer settings, and the product identifiers you want to look up (GTIN, EAN, UPC or ASIN, depending on the source). No proxies, scrapers or marketplace accounts are needed on your side. Get API Key.
Each result is a normalized record: the resolved product, the observed offers with price, shipping, currency, condition and availability, the seller identity behind each offer, an observed_at timestamp, and a confidence score. The same fields appear on every source, so you do not write marketplace-specific parsing. Architecture and data model.
No. kwery runs the collection layer: access, product matching, retrieval, parsing, normalization and validation. You call one API and receive structured records, and when a marketplace changes its site the change is absorbed centrally rather than in your code. Managed data vs web scraping.
The first successful call typically takes minutes using the copy-paste examples in the docs. Teams going to production usually spend a few days wiring webhook delivery, retries and their own storage around it. API documentation.
You get an explicit not-found result rather than an empty or guessed record. That answer is cached for 10 days by default (adjustable per request with not_found_max_age), and re-checks inside that window are served from cache and cost 0 credits. After the window expires, the next check triggers a fresh crawl, which is billed at the source's normal credit rate. Pricing and credits.
Yes. Every endpoint can be exercised against the built-in mock server in the documentation before you sign up. It returns correctly shaped example responses, accepts any dummy key, and consumes no credits, so you can build and test an integration end to end and then swap in the live base URL and a real key. Try it, no account needed.
Request API access and the team will walk through sources, volumes, and integration in detail.