New: try datawebot in the playground, no sign-in needed →

ECOMMERCE MCP SERVER · API · CLI

The world's ecommerce.
One MCP.

search in any country, in its own language
every seller's price for a product
store catalogues, as structured data

200 free credits. No card, no subscription.

Give this to your agent

set up https://mcp.datawebot.com/setup.md

Or connect it yourself

Using Claude.ai or ChatGPT? Add it as a connector →

datawebot

Brooks Revel 8 Running$84.97★ 4.6

The Home Depot
AliExpress
Zalando
IKEA
Allegro
Otto

datawebot is an ecommerce MCP server and API that lets AI agents search products in online shops worldwide, compare every seller's price for a product, and extract Amazon, Walmart and brand-store data as structured records.

Real results

Ask in plain words. Get products from real shops.

Find running shoes under $100 that I can buy in the US. Show them as cards with price and shop.

search_products({"query":"running shoes under $100","country":"US"})

adidas Men's Adizero Pacer Running Shoes$80.00 · adidas
Brooks Men's Revel 8 Running$84.97 · DICK'S Sporting Goods
Merrell Men's Agility Trail$49.00 · Sierra
Nike Men's Revolution 8 Running Shoe$52.49 · Famous Footwear

I need a 4 person camping tent from a UK shop. Show a few options with prices.

search_products({"query":"4 person camping tent","country":"GB"})

Eurohike Sendero 4 Family Tent£100.00 · Amazon.co.uk
Pro Action 4 Person 1 Room Tunnel Camping Tent£80.00 · Argos
Halfords 4 Person Vis A Vis Tent XL£50.00 · Halfords
Decathlon 4 Man Tent£119.99 · Decathlon UK

Search Japanese shops for an electric toothbrush, the brush itself and not replacement heads (電動歯ブラシ本体), and show prices in yen.

search_products({"query":"電動歯ブラシ本体","country":"JP"})

ライオン電動アシストブラシ電池式歯ブラシ本体 +フロス付き¥2,078 · Amazon公式サイト
Braun ブラウンオーラルB プラックコントロール DB5010N¥1,728 · Yahoo!ショッピング - サンドラッグe-shop
本体ライオン¥1,980 · カインズ
ブラウン BRAUN 電動歯ブラシ Oral-B プロ600 ブラック D165131UBK¥6,980 · ソフマップ・ドットコム

Real result · retrieved 23 Oct 2026 · See all prompts

Works where you build

In your agent today. In your product tomorrow: MCP, REST API and CLI.

Same tools, two ways in. No new SDK to learn.

Agent session

You    I need a 4 person camping tent from a UK shop. Show a few options with prices.
Agent  search_products {"query":"4 person camping tent","country":"GB"}
       → Eurohike Sendero 4 Family Tent · £100.00 · Amazon.co.uk
       → Pro Action 4 Person 1 Room Tunnel Camping Tent · £80.00 · Argos
       → Halfords 4 Person Vis A Vis Tent XL · £50.00 · Halfords
       → Decathlon 4 Man Tent · £119.99 · Decathlon UK
       (retrieved 23 Oct 2026)

Connect once with the setup prompt above. Your agent calls the tools when your request needs them.

Python

# pip install mcp
import asyncio, os
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client

async def main():
    headers = {"Authorization": f"Bearer {os.environ['DATAWEBOT_API_KEY']}"}
    async with streamablehttp_client("https://mcp.datawebot.com/mcp", headers=headers) as (read, write, _):
        async with ClientSession(read, write) as session:
            await session.initialize()
            result = await session.call_tool("search_products", {"query":"4 person camping tent","country":"GB"})
            print(result.content[0].text)

asyncio.run(main())

Create a key on the dashboard. LangChain, the OpenAI Agents SDK, the Vercel AI SDK and n8n work the same way: point their MCP client at the server with your key.

curl

curl https://api.datawebot.com/v1/jobs \
  -H "Authorization: Bearer $DATAWEBOT_API_KEY" \
  -H "Idempotency-Key: amazon-coffee-0001" \
  -H "Content-Type: application/json" \
  -d '{"blueprint_id":"amazon.search_items","url":"https://www.amazon.com/s?k=coffee","parameters":{"max_pages":1}}'

# Then poll GET /v1/jobs/{job_id} until it finishes,
# and read GET /v1/jobs/{job_id}/results, following next_cursor.

One Job is one URL. See the API reference for statuses, errors and limits.

Endpoints

More than 28,000 stores. One product record.

Amazon in 20 countries, eBay in 17 and Zalando in 17, plus brands like Reebok, Steve Madden and Skechers in their country stores. Every search and product answer comes back in the same structured product record (title, price and currency, availability, link), so your agent never parses a page.

Search endpoints

Search result pages for your keyword on Amazon, eBay, Walmart and Zalando; brand stores return up to 10 matching products.

Product endpoints

One product page in full, with options, prices and availability.

Catalogue endpoints

A brand store's products with variants and prices, up to 250 a Job, ready for CSV. 5 credits a page; failed Jobs are free.

sites: 28,510
endpoints: 85,387
countries: 113

Marketplaces and brands, in every country we cover

Amazon

20 countries

Search, product

eBay

17 countries

Search, product

Zalando

17 countries

Search, product

Reebok

6 countries

Search, product, catalogue

Steve Madden

9 countries

Search, product, catalogue

Skechers

8 countries

Search, product, catalogue

New Era

8 countries

Search, product, catalogue

Decathlon

6 countries

Search, product, catalogue

BURGA

14 countries

Search, product, catalogue

EcoFlow

12 countries

Search, product, catalogue

Jackery

5 countries

Search, product, catalogue

Brandy Melville

5 countries

Search, product, catalogue

See every store Jobs API reference

Search the stores your agent can extract

Type a store name or domain, e.g. amazon.de.

Browse the store catalogue

How it works

How it works: connect, approve, ask.

Connect your coding agent, approve access, then ask for the product data you need.

01
Give your agent the setup prompt.
Use a coding agent with a terminal and remote MCP support. The setup instructions guide the connection.
02
Approve access in your inbox.
Check the sign-in code and approve your agent. Sign-up is free and open to anyone; no API keys pasted into chat.
03
Ask. Inspect. Put it to work.
Ask for products across online shops, compare seller offers with direct links, and keep the retrieval time with each result.

Search results are best-effort snapshots, not verified live prices. Amazon in 20 countries, Walmart and brand stores run as extraction Jobs, open to every account with 200 free trial credits. See every supported store.

What your agent can do. Search, compare, extract.

Product search API

Search products by keyword on one shop or across shops, in any country and its language: title, price, shop, rating and a product_id for each.

Product data API with every seller's price

Read one product's details and, for a product sold by several shops, each seller's price, total with shipping, stock and a direct link.

Ecommerce scraper API

Run extraction Jobs on search, product and catalogue pages from Amazon, Walmart and brand stores, with every store listed in the directory.

Pricing

Free to try. Pay per request.

To start: 200 creditsFree for every account, no card needed.
Per request: 5 creditsFor a search. A product read, with every seller's price, is 1 credit. Failed requests are free.
After that: From $10Credit packs in US dollars, with no subscription. Credits don't expire.

See pricing

FAQ

Questions about datawebot

What is datawebot?

Is there a free plan?

Yes. Every account gets 200 free credits, enough for 40 searches or 200 product reads. No card is needed.

Which agents can I connect?

Claude Code, Codex, Cursor, opencode, ChatGPT, Claude.ai and any MCP client. Coding agents take the setup prompt; chat apps add https://mcp.datawebot.com/mcp as a connector. From your own code, use any MCP client with an API key.

Which shops does search cover?

Online shops worldwide, through product search in the country you choose. Coverage differs by country and product.

Which sites can my agent extract?

Amazon in 20 countries, Walmart and brand stores: more than 28,000 online stores in all. Every one is listed in the store directory.

Do I need an API key?

Not in your agent: sign in by email, free and open to anyone. For your own code, create a key on the dashboard.

What does it cost?

Every account starts with 200 free credits. A search costs 5 credits and a product read 1, with every seller's price. Failed requests are free; a page that loads but has no results is still charged. Then credit packs from $10, with no subscription.

Does datawebot track prices over time?

No. Each call is a dated snapshot; your agent or app keeps the history.

Are sponsored results included in search?

Yes, but kept apart from the organic results. Listings the shop or search marks as sponsored come back in a separate sponsored list, each flagged as sponsored, and agents are told to label them. A sponsored shop offer is flagged and listed last. When your query includes a model code, such as WH-1000XM5, results for other models are left out; if nothing matches, all results are shown and the summary says so.

Do the logos mean every source is supported?

No. The logos illustrate ecommerce breadth, not partnerships or live integrations. Product search returns best-effort listings from online shops. Extraction covers the sites in the endpoint catalogue.

ECOMMERCE MCP SERVER · API · CLI

Ask in plain words. Get products from real shops.

In your agent today. In your product tomorrow: MCP, REST API and CLI.

More than 28,000 stores. One product record.

Marketplaces and brands, in every country we cover

Search the stores your agent can extract

How it works: connect, approve, ask.

Give your agent the setup prompt.

Approve access in your inbox.

Ask. Inspect. Put it to work.

What your agent can do. Search, compare, extract.

Product search API

Product data API with every seller's price

Ecommerce scraper API

Free to try. Pay per request.

Questions about datawebot