What 30book does
The 30book MCP server connects an MCP client to 30book.com, an Iranian online bookstore listing printed books in Persian and other languages. It is intended for catalog research rather than transactions. Agents can search titles, inspect editions, compare publishers, check Toman prices and discounts, review availability, and examine bibliographic details.
The 30book MCP server is part of the E-Commerce category, but its access is strictly read-only. It does not log in, add items to a cart, place orders, manage wishlists, or submit reviews. It also does not cover e-books or audiobooks.
How it works
The server retrieves and parses pages from 30book.com without using a public 30book API. Book, home, and landing pages are read from server-rendered HTML. Search and catalog pages use a Blazor Server session over SignalR long polling; the server processes the returned render data and extracts book cards without simulating browser clicks.
Search and listing requests support paging, sorting, and filters. A typical investigation starts with a title search, continues to a detailed book lookup, and then branches to similar books or reader reviews. For an author or publisher, resolve the name to an identifier first, then pass that identifier to the relevant catalog tool.
Responses are compact JSON. Prices are reported in Toman, with separate list and final-price fields plus a discount percentage. Stock is represented as yes or no, not as an inventory count. List requests generally take two to five seconds, while individual book pages usually take about one second.
Setup and configuration
Install Python 3.10 or newer and uv, then run the server with:
Claude Code can register it with:
claude mcp add 30book -- uvx 30book-mcp
The server uses stdio transport. Claude Desktop, Cursor, and VS Code can run uvx with 30book-mcp as its argument. Installing the package with pip install 30book-mcp is another supported option.
No login, account, or API key is required. If 30book.com cannot be reached from the host network, set SIBOOK_MCP_PROXY to an HTTP proxy URL. The server intentionally does not use general system proxy variables for site requests.
Tools and capabilities
The 14 tools use the sb_ prefix. Search and discovery tools include:
sb_search for title searches, sorting, stock, publisher, author, and translator filters, with 20 results per page.
sb_browse for subjects, sections, topics, fiction genres, national literatures, awards, and tags.
sb_publisher and sb_author for publisher and person catalogs.
sb_find_publisher, sb_find_person, sb_categories, and sb_collections for resolving identifiers and browsing available classifications.
Book-level tools include sb_book, sb_similar, and sb_reviews. A book record can include prices, discount, stock, authors, translators, publisher, page count, cover type, dimensions, ISBN, print run, publication year, categories, rating, and description. Popularity and pricing tools include sb_bestsellers, sb_new_releases, and sb_deals.
Limitations and notes
The 30book MCP server works best with Persian search terms for titles and names. Some names may use Arabic character variants, so a shorter query or the extended search option may help when results are empty. Keep request rates low because the project is unofficial and is not affiliated with 30book.
Use the tool identifiers returned by lookup operations when calling author, publisher, browse, or book tools. The service provides availability as a boolean-style value rather than a stock quantity, and site reachability may depend on the network where the MCP process runs.