MCP Server

The Easyparser MCP server exposes the Real-Time API operations to AI agents as callable tools through the Model Context Protocol. You connect your MCP client to one endpoint, and your agent can fetch Amazon product, offer, search and seller data, monitor bulk jobs and check your credit balance without custom integration code. Coding agents can also use it to test requests while they build your integration (see Build an integration with an AI coding agent).

The MCP server is a thin, read-only adapter over the Real-Time API. It calls the same operations, so credit costs, marketplaces and data shapes match the Real-Time API pages. This page covers connecting a client and what each tool does.

Before you begin

You need an Easyparser API key. Sign up at https://app.easyparser.com/signup (the free plan includes 100 credits per month and needs no credit card), then copy your key from Account > Plan at https://app.easyparser.com/account/plan.

Wherever a setup step below shows YOUR_API_KEY, paste your key.

Note

The list_operations tool works without an API key, so an agent can explore the operation catalog before you sign up.

Available tools

The server provides 18 tools. Data tools cost the same credits as the matching Real-Time API operation; monitoring and account tools are free. See Credit Consumption for the full rules.

• Each tool description tells the agent what the tool returns, when to use it, when a cheaper tool fits better, and what it costs.
• Every data tool response includes success, operation, data, credits_used, credits_remaining and request_id, so the agent can track its budget.
• Rate-limit and authentication errors come back as readable guidance instead of raw HTTP status codes.

Product data

Tool	Operation	Returns	Credits
get_product_detail	DETAIL	Full product profile: title, brand, images, Buy Box price, rating, BSR, variants, top reviews	1, or 2 with A+ content
get_product_offers	OFFER	Seller offers: price, condition, FBA or FBM, seller rating, Buy Box winner	1 per page, max 5 pages
get_bestseller_rank	BEST_SELLERS_RANK	Current Best Sellers Rank by category	1
get_package_dimensions	PACKAGE_DIMENSION	Package dimensions, weight and Amazon fee category	1
get_similar_products	SIMILAR_PRODUCTS	Similar products for an ASIN: title, brand, category, image	1, 6 or 11 for 10, 50 or 100 results
get_sales_history	SALES_ANALYSIS_HISTORY	Weekly trends for up to 12 months: views, sales, price, BSR	5 plus 1 per month of history, max 17

Search and discovery

Tool	Operation	Returns	Credits
search_products	SEARCH	Keyword or search URL results with price, rating, Prime status and badges	1 per page, max 5 pages
lookup_product	PRODUCT_LOOKUP	ASIN for a UPC, EAN, GTIN, ISBN, JAN or MINSAN code, with a product snapshot	1

Sellers

Tool	Operation	Returns	Credits
get_seller_profile	SELLER_PROFILE	Legal business name, business address, feedback summary for the last 1, 3 and 12 months and all time	1
get_seller_products	SELLER_PRODUCTS	A seller's catalog with brand, price, Prime status and ratings	1 per page, max 5 pages
get_seller_feedback	SELLER_FEEDBACK	Individual buyer feedback, filterable by rating and time period	1 per page, max 5 pages

Bulk job monitoring

These tools track jobs you started with the Bulk API or on the Bulk Requests page of the web app (https://app.easyparser.com/bulk-requests).

Tool	Returns	Credits
list_bulk_jobs	Your bulk jobs with status and completed, failed and invalid counts	Free
get_bulk_job_items	Per-item outcomes inside a job, for debugging failed or invalid items	Free
get_bulk_item_result	The parsed data one item produced (kept for up to 24 hours)	Free
get_bulk_webhook_logs	Webhook delivery logs for your bulk items	Free

Error logs

Tool	Returns	Credits
get_error_logs	Real-Time API and Bulk API request errors: error code, channel, operation, request parameters, timestamp	Free

get_error_logs returns the same data as the Errors page of the web app (https://app.easyparser.com/errors).

Account

Tool	Returns	Credits
list_operations	Operation catalog with parameters, credit costs and supported domains. Works without an API key	Free
check_credits	Plan, credits used and remaining, reset date, per-minute credit limits, address usage, daily usage this month	Free

Prompts and resources

Clients that support MCP prompts can start these workflows in one step:

Prompt	What it does
analyze_competitor_product	Listing details, seller offers and 3-month sales and BSR trends for one ASIN
estimate_fba_logistics	Package dimensions, weight and fee category for FBA planning
match_barcodes_to_asins	Converts a list of UPC, EAN, GTIN or ISBN codes into ASINs
vet_amazon_seller	Legal identity, reputation trend and negative feedback for a seller
product_market_research	Top organic results and demand comparison for a keyword niche

The server also provides free, read-only resources that work without an API key:

Resource	Contents
easyparser://operations	Every operation with its credit cost
easyparser://marketplaces	The Amazon domains the MCP server accepts
easyparser://guide	Rules for picking tools and saving credits

Connect your client

You can connect in two ways:

Option	Endpoint or command	Use it when
Hosted remote server	https://mcp.easyparser.com/mcp	Your client supports remote (HTTP) MCP servers. No install needed.
Local server	npx -y easyparser-mcp	Your client supports only local stdio servers, or you want to run the server yourself.

All hosted setups use the same endpoint, https://mcp.easyparser.com/mcp, and send your key in the Authorization: Bearer YOUR_API_KEY header.

Caution

Authorization: Bearer works only for the MCP server. The REST APIs use the api_key query parameter (Real-Time API, Account API) or the api-key header (Bulk API, Data Service).

Cursor

The GitHub repository has one-click install links for the remote server and the local (npx) server. They open Cursor with the configuration filled in; replace YOUR_API_KEY when prompted.

To add the server manually, edit .cursor/mcp.json:

.cursor/mcp.json

{
  "mcpServers": {
    "easyparser": {
      "type": "http",
      "url": "https://mcp.easyparser.com/mcp",
      "headers": { "Authorization": "Bearer YOUR_API_KEY" }
    }
  }
}

VS Code

The GitHub repository has a one-click install link for the remote server. VS Code then asks for your API key in an input prompt.

To add the server manually, edit .vscode/mcp.json:

.vscode/mcp.json

{
  "servers": {
    "easyparser": {
      "type": "http",
      "url": "https://mcp.easyparser.com/mcp",
      "headers": { "Authorization": "Bearer YOUR_API_KEY" }
    }
  }
}

Claude Code

Run this command in your terminal:

claude mcp add --scope user --transport http easyparser https://mcp.easyparser.com/mcp \
  --header "Authorization: Bearer YOUR_API_KEY"

Claude Desktop and claude.ai

Add the server as a custom connector:

1. Open the connector form

   Go to Settings > Connectors > Add > Add custom connector.

2. Enter the server details

   Set Name to Easyparser and Remote MCP server URL to https://mcp.easyparser.com/mcp, then click Continue.

3. Set authentication

   Under Authentication, select None. Easyparser uses an API key, not OAuth.

4. Add the API key header

   Under Request headers, add a header named authorization with the value Bearer YOUR_API_KEY and mark it Required. Click Add.

Manus

Go to Settings > Integrations > Custom MCP Servers > Add Server and fill in the form:

Field	Value
Server name	Easyparser
Transport type	HTTP
Server URL	https://mcp.easyparser.com/mcp
Custom headers	Name Authorization, value Bearer YOUR_API_KEY

Click Test to check the connection, then Save.

Any other MCP client

The hosted endpoint uses the Streamable HTTP transport at https://mcp.easyparser.com/mcp. Send your key in one of these headers:

Header	Value
Authorization	Bearer YOUR_API_KEY (recommended)
x-api-key	YOUR_API_KEY
api-key	YOUR_API_KEY
apikey	YOUR_API_KEY

Local stdio server

For clients that only run local servers, start the npm package easyparser-mcp with your key in the EASYPARSER_API_KEY environment variable. A typical client configuration:

mcp.json

{
  "mcpServers": {
    "easyparser": {
      "command": "npx",
      "args": ["-y", "easyparser-mcp"],
      "env": { "EASYPARSER_API_KEY": "YOUR_API_KEY" }
    }
  }
}

The package needs Node.js 18 or later.

Self-host the HTTP server

You can run the HTTP server on your own infrastructure. Install the package globally and start easyparser-mcp-http:

npm install -g easyparser-mcp
PORT=3000 easyparser-mcp-http

In PowerShell, set the port with $env:PORT="3000"; easyparser-mcp-http. PORT defaults to 3000. From a clone of the repository, you can also run node dist/http.js or build the Docker image:

docker build -t easyparser-mcp .
docker run -p 3000:3000 easyparser-mcp

The server exposes these endpoints:

Endpoint	Purpose
POST /mcp	MCP endpoint (Streamable HTTP)
GET /health	Health check

Clients authenticate with the same headers as the hosted server. If a request has no key header, the server falls back to the EASYPARSER_API_KEY environment variable.

Using it in requests

After you connect, ask your agent in plain language. The agent picks the tools, and each call is charged like the matching Real-Time API request.

What is the current price and rating of ASIN B0F25371FH?

Find the ASIN for UPC 041604458187, then show me its 6-month sales trend.

Who is the seller behind this product, what is their legal business name,
and what do their negative reviews say?

Compare the top 10 organic results for "stainless steel water bottle" on
amazon.de: prices, ratings, and which ones are Prime.

Is the bulk job I started this morning finished? If any items failed,
tell me which ASINs and why.

Danger

Agents can chain several paid tools for one prompt (for example, a lookup followed by a sales history call). Ask the agent to run check_credits first, or watch credits_remaining in tool responses, to keep usage under control.

Build an integration with an AI coding agent

The MCP server also helps while you write code. When your coding agent (for example Claude Code, Cursor or an agent in VS Code) has the Easyparser MCP server connected, it can call the operations you plan to use, read real responses, and then write your integration against the actual field names. It can run your tests and fix mismatches before you review the code.

1. Connect the MCP server in your coding agent

   Add the server as described in Connect your client, for example with claude mcp add for Claude Code. Give your application its own key in an environment variable, such as EASYPARSER_API_KEY, instead of pasting the key into code.

2. Describe the task and ask the agent to test first

   Name the data you need and tell the agent to try the request through MCP before it writes code:

   Add current Amazon prices from Easyparser to my app. Use the Easyparser MCP
   server to test a DETAIL request for ASIN B0CJB6V2L5 on amazon.com first, then
   write a getPrice(asin) function that calls the Real-Time API with the key in
   EASYPARSER_API_KEY, and add a test with a mocked API response.

3. Let the agent explore, then write the code

   The agent usually calls list_operations (free) to choose the operation and its parameters, then one paid tool, for example get_product_detail (1 credit), to see a real response. It then writes code that calls https://realtime.easyparser.com/v1/request directly with field paths such as result.detail.buybox_winner.price.value, and runs your tests.

Note

MCP tool results contain the API response's result object (returned as data) plus the credit fields. Your application receives the full Real-Time API response, with request_info, request_parameters, request_metadata and result (see the response envelope). Ask the agent to read the data from result and to check request_info.success in your code.

Caution

Every paid tool call the agent makes while testing uses credits. Ask it to run check_credits first, and mock the API in your automated tests so a test run does not send paid requests.

Limitations

These are limits of the MCP tools, not of the API. Use the Real-Time API directly when you need the full parameter set.

• The server is read-only. It cannot create bulk jobs, manage addresses or change your account.
• get_product_offers, search_products, get_seller_products and get_seller_feedback cap max_page at 5 (min_page and max_page both default to 1), so a call can never reach page 6 or later. The Real-Time API accepts later page windows of up to 5 pages.
• get_product_offers accepts one condition filter per call. The Real-Time API lets you combine several condition flags.
• get_seller_feedback takes a seller_id only. The Real-Time API also accepts a seller URL.
• The tools do not expose address_id, the inline address fields (zip_code, city, district, country_code), parameters or cookie, so you cannot request address-specific results through MCP.
• get_bulk_item_result returns results for up to 24 hours. Results can be removed earlier.
• The marketplace list the MCP server accepts may differ from the list in Supported Marketplaces. Read the easyparser://marketplaces resource for the exact list.

Related

• Real-Time API overview
• Credit Consumption
• Supported Marketplaces
• Bulk API overview
• Account API
• n8n