Klook MCP Server

Give your AI assistant live access to Klook — the travel-experiences marketplace. This Actor runs as a Model Context Protocol (MCP) server so Claude, ChatGPT, Cursor, n8n and custom agents can search activities, browse cities, resolve destinations and pull categories on demand. It also works as a classic one-shot scraper if you just want a dataset.

Backed by Klook's mobile API (with in-Python request signing and DataDome handling) for reliable, structured results.

Who it's for

• AI-agent & app builders — add real Klook activity search to an assistant or a RAG/travel-planning workflow without touching anti-bot or auth.
• Travel-tech developers — a hosted MCP endpoint for activity data; no scraping code to maintain.
• Analysts & researchers — interactively explore activities, prices and demand by city from a chat interface.

MCP tools

Tool	What it does
search_klook_activities	Keyword search → activities with price, rating, reviews, bookings, city, GPS, image.
get_klook_city_activities	Browse a city's full activity catalog by numeric city ID.
search_klook_suggestions	Autocomplete cities (with IDs) and activities for a term — resolve destination names.
get_klook_city_info	City overview: travel tips, weather, hot activities, nearby cities.
get_klook_activity_images	Image gallery for a specific activity.
get_klook_categories	Klook's top-level activity category / vertical tree.

Connect it to an AI assistant

1. Start this Actor in Standby mode (it launches an MCP server).
2. Point your MCP client at the server URL + /mcp (Streamable HTTP transport).

Example MCP client config:

{
    "mcpServers": {
        "klook": {
            "url": "https://<your-actor-standby-url>/mcp"
        }
    }
}

Then ask your assistant things like "Find the top-rated Universal Studios Japan tickets on Klook" or "What are the most-booked activities in Tokyo?".

Classic scraper mode

Prefer a dataset? Run it normally with input:

{
    "searchQueries": ["universal studios japan"],
    "cityIds": ["28"],
    "sortBy": "most_reviewed",
    "maxItems": 50
}

Each activity is pushed as a structured row (same fields as the tool output): activity_id, title, activity_url, rating, review_count, booked_count, sell_price, price_from/price_to, city_name, country_name, latitude, longitude, image_url, scraped_at, and more.

Scheduling & integrations

Schedule classic-mode runs and pipe results to Google Sheets, Make, Zapier, S3 or your API via Apify webhooks and the dataset API (JSON, CSV, Excel).

FAQ

What is MCP? The Model Context Protocol — an open standard that lets AI assistants call external tools. This Actor is a hosted MCP server for Klook.

Where does the data come from? Klook's official mobile backend API — the same source the Klook app uses.

How fresh is it? Live at call time.

Do I need the sibling scraper? No — this Actor also scrapes to a dataset. Use the Klook Activities & Experiences Scraper if you only want bulk dataset runs without MCP.

Is it legal? You are responsible for your use of the data; it collects publicly available catalog information. Respect Klook's terms and applicable laws.