tabelog
Provides access to Tabelog's restaurant data, enabling searches by area, genre, budget, and other filters, and retrieving details, reviews, menus, ratings, photos, and vacancy information.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@tabelog@tabelog search for ramen in Sannomiya"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
tabelog-cli
A Tabelog CLI and MCP server. Tabelog has no public API and the Japanese site
sits behind a Cloudflare challenge that rejects anything but a real browser,
but the inbound locales (tabelog.com/en, /kr, /tw, /th) are served as
plain server-rendered HTML with schema.org JSON-LD. This reads those. No
browser, no headless Chrome, no runtime dependencies.
What it reads: restaurant search with Tabelog's own filters (area, genre, keyword, budget band, online-bookable at a date/time/party size) plus two filters the site does not offer (distance from coordinates, open at a given time); a restaurant page with structured weekly hours; reviews; the posted menu; the rating breakdown; photos; and the online-booking calendar with time slots. Reservation itself stays in the browser.
Install
Requires Bun.
bun install
bun link # registers the global `tabelog` commandRelated MCP server: grubhub-mcp
Usage
# top ramen around Sannomiya station, Tabelog score order
tabelog search --area Sannomiya --genre ramen
# Japanese works for both filters and free text; --page walks 20 at a time
tabelog search --area 三宮 --keyword 焼鳥 --page 2
# other orderings the inbound site offers
tabelog search --area Namba --genre okonomiyaki --sort access # most viewed by overseas visitors
tabelog search --area Namba --genre okonomiyaki --sort reserved # most reserved
# budget per person (dinner by default), snapped to Tabelog's bands
tabelog search --area Sannomiya --genre ramen --budget-max 999
tabelog search --area Ginza --genre sushi --budget-meal lunch --budget-min 5000 --budget-max 9999
# only places with an online-bookable table then
tabelog search --area Sapporo --genre sushi --vacancy-date 2026-09-11 --vacancy-time 19:00 --vacancy-people 1
# nearest first from coordinates, within 400 m, and open at that time in Japan.
# both read every result's page (about 2 s for a page of 20)
tabelog search --area Sannomiya --genre ramen --near 34.6946,135.1955 --radius-m 400 --open-at "2026-09-11 14:30"
tabelog search --area Susukino --open-at now
# a restaurant page, by URL or by the id shown in search results
tabelog detail https://tabelog.com/hyogo/A2801/A280101/28043837/
tabelog detail 28043837
# reviews, newest visit first, lunch only
tabelog review 28043837 --by-visit --use-type lunch
# posted menu (food by default; lunch, drink), rating breakdown, photos
tabelog menu 28002413
tabelog menu 28002413 --kind drink
tabelog rating 27000401
tabelog photo 27000401 --mode owner
# online-booking calendar and time slots for a date and party size
tabelog vacancy 27000401 --date 2026-09-11 --people 1 --time 19:00
# what does a word resolve to?
tabelog suggest Sannomiya
# Korean page text (names, categories, transport); filters resolve the same way
tabelog search --area Namba --genre okonomiyaki --locale kr
TABELOG_LOCALE=kr tabelog detail 28043837
# machine-readable output for any command
tabelog search --area Sannomiya --genre ramen --jsonArea and genre names are matched against the suggest index, which knows English
and Japanese names only. A Korean area name returns nothing; use the English or
Japanese one and switch --locale kr for Korean output. Genres are pickier
than areas: the index has Tabelog's own labels, so 串カツ resolves where
kushikatsu does not. A genre that fails to resolve is searched as a keyword
instead, and the result header says so. Landmarks (Dotonbori) are not areas;
use the nearest station.
Times are Japan time. --open-at reads the English page's weekly hours; a
restaurant with no parsable hours is kept and marked unknown, never dropped.
--near is straight-line distance and applies to the fetched page only, so
pair it with an area that already narrows the candidates.
MCP
tabelog mcpServes the same commands as MCP tools over stdio: search, detail,
review, menu, rating, photo, vacancy, suggest. The server is a
hand-written JSON-RPC loop rather than the official SDK, which pulls in a
hundred packages for HTTP transports this never uses. Register it with your
host as command tabelog, arguments ["mcp"].
How it reads the site
Page | Source |
Search |
|
Area / genre resolution |
|
Detail | schema.org |
Reviews |
|
Menu |
|
Rating |
|
Photos |
|
Vacancy |
|
Distance / open now | Computed here from each result's JSON-LD coordinates and parsed hours; Tabelog's inbound site has neither filter. |
Id to URL |
|
If Tabelog changes its markup the parsers here will need to follow. Every selector is a BEM class name the site has kept stable for years, and each command fails loudly rather than returning an empty page as success.
License
MIT
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
No tool schema history has been recorded yet.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Live Google Maps business search, review, and photo data for AI agents over MCP.
17+ Japan MCP tools (weather/calendar v2/local-pack/enrich). x402 on Base, wallet-free trial.
Pay-per-use tool marketplace for AI agents. Search, price-check, and call APIs via MCP.
All HasData scraping tools in one MCP server: Google, TikTok, Instagram, maps, e-commerce and more.
Related MCP Servers
- FlicenseAqualityDmaintenanceMCP server for Palate Network — query trust-weighted venue recommendations from an AI agent network. 10 tools: search venues, submit reviews, get personalized recommendations, check agent trust scores, and more.1117-
- AlicenseNot gradedqualityCmaintenanceProvides programmatic access to Grubhub's food delivery platform, enabling restaurant search, menu browsing, cart management, order placement, and delivery tracking through MCP tools.MIT
- AlicenseNot gradedqualityCmaintenanceMCP server that transforms Gurunavi into a browser-driven semantic proxy, providing restaurant search, details, and reservation management through normalized interfaces with safety controls.MIT
- AlicenseAqualityDmaintenanceEnables AI agents to search restaurants, check availability, and book reservations on OpenTable, with persistent sessions and MFA handling.9461MIT
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/krisamin/tabelog-cli'
If you have feedback or need assistance with the MCP directory API, please join our Discord server