For shopping assistants
Give your shopping assistant real prices
The ShopSavvy Data API gives shopping assistants and agents product search, current prices from tens of thousands of retailers sorted by price, price history back to 2011, dense from 2019, and a deals feed. Call it over REST, or connect an assistant to the ShopSavvy MCP server by URL with no key.
Search products, compare current prices across tens of thousands of retailers, and surface deals: the product and pricing layer behind shopping assistants and agents. Available as a REST API or a one-command assistant connector.
Starter $49/mo · 1,000 credits · card required, cancel before the trial ends and pay nothing
- 30M+ productsin the catalog
- 60M+ offerscurrent, each with the time it was last seen
- Since 2011price history, dense from 2019
- 10 SDKsopen source, in 10 languages
Measured against the ShopSavvy production database, as of Oct 11, 2026.
What can you build with it?
Find-me-the-best-price
Let your assistant answer "where’s this cheapest?" with a comparison across tens of thousands of retailers, sorted by price.
Conversational product search
Turn a natural-language request into ranked product results your assistant can talk about, then drill into prices and reviews.
Deal discovery
Surface current, community-voted deals so your assistant can flag a good price.
How does it work?
- Step 1
Find the product
GET /v1/products/search?q= turns the user’s words into product candidates (1 credit per product returned).
- Step 2
Compare prices
GET /v1/products/offers for the chosen product returns every offer with retailer, price, condition, availability and link.
- Step 3
Add context
GET /v1/products/offers/history answers "is this a good price?", GET /v1/products/reviews adds a TLDR review, and GET /v1/deals surfaces current deals.
Which response fields will you use?
| offers[].price, retailer, URL | the answer and where to buy |
|---|---|
| offers[].availability, condition | only recommend what can be bought new |
| offers[].history[] | "is this a good price?" answers |
| title_short, images[] | compact product cards in chat |
Which identifiers match best?
Assistants usually start from words, so keyword search comes first; follow-up calls should use the barcode or ASIN from the search result for an exact match.
How fresh is the data?
Every offer carries a timestamp of when ShopSavvy last saw that price. Products you schedule are refreshed hourly, daily or weekly for retailers we can re-check through an official retailer API; other offers update as ShopSavvy sees them.
What does the first call look like?
Copy-paste in your language of choice. Open-source SDKs in 10 languages (Python, TypeScript, Go, Rust, Java, Kotlin, Swift, PHP, Ruby, C#) wrap the same calls.
curl -X GET 'https://api.shopsavvy.com/v1/products/search?q=sony%20headphones' \
-H 'Authorization: Bearer YOUR_API_KEY'What would it cost?
An example workload, not a customer: 250 price comparisons a month for products the assistant has already identified.
| Reading offers (all retailers) | 250 products × 1 read × 3 credits | 750 |
|---|---|---|
| Credits a month | 750 |
Fits Starter at $49/month (1,000 credits). Misses cost 0, so real usage is the same or lower.
When is it not the right fit?
- You need prices in currencies other than US dollars: offers are returned in US dollars.
- You need checkout or ordering. The API returns product and price data and links to each retailer; it does not place orders.
Just want your own assistant (Claude, ChatGPT, Cursor) to look up prices? Connect the MCP server, free without a key →
Which tools does it plug into?
Questions
Is there a connector for assistants?
Yes. The ShopSavvy MCP server exposes 15 MCP tools for product lookup, prices, history and deals. It is free without a key (10 tool calls a minute, 100 an hour); an API key raises it to 100 a minute, and tool calls then use plan credits. The same data is available over plain REST.
Do you sort results by price?
Always. Results are sorted by price, and no retailer can pay to rank above another.
Can my assistant search by keyword?
Yes. The search endpoint takes free text and returns matching products at 1 credit per product returned; follow up with the offers endpoint for prices.
Can it tell users whether a price is good?
Yes. The history endpoint returns dated prices for any window of up to 366 days, from an archive with price history back to 2011, dense from 2019, so the assistant can compare today’s price with recent lows.
What would this cost us?
It depends on volume, and you only pay for results returned. For example, 250 price comparisons a month for products the assistant has already identified. That is 250 products × 1 read × 3 credits = 750: 750 credits a month, which fits the Starter plan ($49 a month, 1,000 credits).
Is there a trial?
Yes. Starter ($49 a month for 1,000 credits) starts with a 14-day trial; the larger plans bill from the first day. Checkout asks for a card and shows the date the trial ends, and cancelling before it ends costs nothing.
Related industries
Build your assistant
Start the 14-day Starter trial: $49/mo for 1,000 credits after the trial, cancel anytime.