https://api.kwery.co/mcpTools
One API key for all of them, so the data and the credits are the same as in the API.
get_product_offersLive price, offers and trust score for one product on one marketplaceCost: Source ratecompare_pricesOne product on up to 5 marketplaces, ranked by priceCost: Sum of sourcessubmit_batchUp to 1,000 products on one marketplace as a background job. Shows the credit estimate first and spends nothing until you confirmCost: Per resultget_batch_statusHow far a batch job has gotCost: Freeget_batch_resultsThe finished results, page by pageCost: Freelist_sourcesMarketplaces, countries, identifier types and credit ratesCost: Freeget_accountYour plan, credit balance and credits usedCost: FreeIdentify products by GTIN, ASIN or the marketplace's own ID. Two ready prompts, price_check and catalog_sweep, appear in clients that show MCP prompts.
Prompts
Setup
Sign up, copy your API key from the dashboard, add the server. The full steps for every client are in the MCP guide.
In the app
On Team and Enterprise plans an owner adds kwery once under Organization settings → Connectors, and the team shares that key and its credits.
Response
The same data as the API, so the assistant can tell whether to trust a price before anyone acts on it.
Lowest price
The cheapest new, in-stock offer including shipping, with suspect offers left out. A second field covers used and open-box offers.
Offers
Every offer with seller, price, shipping and stock.
Trust score
On every result. Below 0.7, a person should check it first.
Not found
When a product isn't listed, the assistant is told so instead of guessing.
{
"tool": "compare_prices",
"gtin": "4548736151757",
"country": "DE",
"lowest_price": {
"source": "idealo",
"price": 279.00,
"shipping": 0,
"currency": "EUR",
"trust_score": 0.94
},
"lowest_price_any_condition": { "source": "amazon", "price": 249.00 },
"results": [
{ "source": "idealo", "offers": 14, "trust_score": 0.94 },
{ "source": "amazon", "offers": 6, "trust_score": 0.97 },
{ "source": "google_shopping", "offers": 9, "trust_score": 0.88 }
]
}Credits
The MCP server uses your normal kwery credits at each marketplace's rate. The free plan works too.
Each lookup costs the same credits as the matching API call, at that marketplace's rate.
Cached results and failed lookups cost 0 credits.
Batches show a credit estimate first and spend nothing until you confirm.
Directories
Use Cases
If the data behind an answer is stale or matched to the wrong product, the assistant recommends the wrong offer.
Assistants quote prices from comparison sites, marketplaces and Google Shopping. Live, GTIN-matched data keeps those answers current.
Agents that check your own catalog against the market need the right product, not a lookalike. Trust scores flag what a person should review.
Anyone with an AI assistant can ask for a price check in plain language, using the same data and credits as the API.
FAQ
Setup, costs and limits for using kwery in AI assistants.
Browse AI Agent FAQsChatGPT, Claude (on the web, Claude Desktop and Claude Code), Gemini, Grok, Perplexity, Cursor and VS Code. ChatGPT and Perplexity connect by signing in to kwery; the others use your API key. Any other app that supports remote MCP with OAuth or a custom header works too.
No. Add kwery as a custom connector and sign in with your kwery account. Usage comes out of that account's credits.
The same credits as the API, at each marketplace's rate. Cached results and failed lookups are free. Free-plan credits work too.
Batches show a credit estimate first and spend nothing until you confirm.
On the free plan, compare_prices is limited to 30 calls per minute. A batch takes up to 1,000 products.
All kwery sources except AliExpress, which is available through the REST API. In Germany, if you don't name any, kwery checks Amazon, idealo and Google Shopping.
Sign up, copy your API key, add the server.