Rakuten Japan MCP

What rakuten-japan-mcp does

The rakuten-japan-mcp MCP server provides agent access to Rakuten Ichiba product and ranking data. It uses Rakuten’s official APIs rather than extracting information from rendered web pages, so the returned data is structured for programmatic use.

The project supports two main data operations:

• Search Rakuten Ichiba products by keyword.
• Retrieve ranking data, including current rankings, genre-specific rankings, male or female rankings, and rankings from a specified historical era.

Product results can contain the item name, price, product URL, shop information, review data, and image information. The repository also packages the same API access as an Apify actor, which can be run from the Apify Console, through the Apify API, or on a schedule.

How it works

The MCP interface exposes three tools. search_rakuten performs keyword-based product searches. search_rakuten_ranking retrieves ranking results and accepts ranking-related options such as genre, ranking type, and date. get_actor_info reports the available tools and pricing information.

Search requests use Rakuten’s Ichiba item search API. Ranking requests use Rakuten’s Ichiba ranking API. Search calls require both a Rakuten application ID and access key, while ranking calls require the application ID. The available ranking response is limited to 30 items per call, matching the stated ranking API limit.

The rakuten-japan-mcp MCP server is suited to assistants that need current marketplace results for research, comparison, product discovery, or ranking analysis. It can be connected to MCP clients including Claude and Cursor through the actor’s standby endpoint.

Setup and configuration

Run the project as an Apify actor or connect an MCP client to its standby endpoint. For actor usage, product search is selected by providing a search keyword, while ranking mode is enabled with the ranking option.

Relevant inputs include:

• searchKeyword for terms such as Japanese product names or categories.
• ranking to switch from product search to ranking retrieval.
• genreId to limit rankings to a Rakuten genre.
• rankingType with item, male, or female as the supported choices.
• era for a historical ranking date in YYYYMMDD format.
• maxResults to set the requested result count, with 30 as the default and ranking results capped at 30 per call.

Configure RAKUTEN_APP_ID for Rakuten API access. Product search also needs RAKUTEN_ACCESS_KEY; ranking requests only require the application ID. The README identifies these as secret environment variables on the Apify actor.

Tools and capabilities

search_rakuten supports keyword product discovery and returns structured marketplace fields including titles, prices, links, shops, reviews, and images.

search_rakuten_ranking supports current rankings, genre filtering, male or female ranking types, and historical ranking queries where the API supports the selected era.

get_actor_info lists the available actor tools and their pricing. The documented pay-per-event rates are $0.001 for a rakuten-search event and $0.001 for a rakuten-ranking event.

The rakuten-japan-mcp MCP server is registered in the official MCP Registry. The repository also documents automated deployment on pushes to the master or main branches, along with pytest and strict mypy commands for project checks.