Crawlora MCP

What Crawlora-org/crawlora-mcp MCP server does

The Crawlora-org/crawlora-mcp MCP server gives MCP clients access to a broad catalog of structured public-web data operations. The supplied tools cover searches and detail lookups across travel, accommodation, flights, activities, commerce, jobs, anime, books, and social platforms. Responses are normalized JSON rather than raw HTML, so an agent can consume fields such as titles, prices, ratings, descriptions, dates, locations, images, and review information according to each tool's schema.

Examples in the catalog include Agoda hotel, home, activity, flight, and airport searches; Airbnb rooms, hosts, listings, calendars, and reviews; Amazon product search, product details, and suggestions; Amazon Jobs search and job details; and AniList anime, character, ranking, schedule, staff, cast, and recommendation queries. The available catalog is broader than these examples and is generated from the project's tool definitions.

How it works

The hosted deployment uses Streamable HTTP at https://mcp.crawlora.net/mcp. A compatible client sends requests to that endpoint with a Crawlora API key, preferably in an Authorization: Bearer ... header; the x-api-key header is also accepted. Missing or invalid credentials produce a 401 response.

For self-hosting, the repository includes a small stdio server that wraps the Crawlora REST API at https://api.crawlora.net/api/v1. The local process exposes the same tool set and forwards requests using the configured API key. Tool definitions are generated from tools.json, with each definition mapping to one REST endpoint and specifying its input schema.

Setup and configuration

The hosted option does not require Node.js. Configure the endpoint and authentication header in an MCP client that supports Streamable HTTP. The README provides configurations for Claude Desktop, Cursor, Windsurf, Cline, Claude Code, Codex CLI, VS Code, Gemini CLI, and generic MCP clients.

The local option requires Node.js 20 or newer when using npx. The directly runnable command is:

CRAWLORA_API_KEY=your-key npx -y crawlora-mcp

A cloned repository can instead be started after npm install with node index.mjs, or built and run through Docker. Keep the key in an environment variable rather than committing it to client configuration or source control. Accounts include 2,000 free credits per month without a credit card, according to the README.

Tools and capabilities

• Search public listings, products, jobs, anime, flights, hotels, homes, activities, and other indexed resources.
• Retrieve normalized detail records for supported products, properties, rooms, hosts, activities, jobs, titles, characters, and books.
• Resolve search inputs such as city or airport names into platform-specific identifiers.
• Retrieve paginated reviews, listings, staff, cast, recommendations, schedules, and other related records where supported.
• Receive clean JSON fields instead of scraping and parsing source HTML in the agent.

The Crawlora-org/crawlora-mcp MCP server is suited to agents that need read-only public information. The supplied material does not describe booking, purchasing, account access, authenticated user data, or write operations.

Limitations and notes

Data comes from public sources and may vary by platform, region, query, and source availability. Some tools require identifiers returned by an earlier search; for example, Agoda flight amenities expects flight segments from a flight-search response, and several detail tools require a platform-specific id or ASIN. Airbnb results and other platform responses are normalized, but the material does not guarantee completeness or freshness beyond the upstream public source.

Use of the hosted or local service requires a Crawlora API key. The README describes a free monthly credit allowance, but it does not specify paid-plan pricing or credit consumption rules. The repository is described as open source, but no license name is provided in the supplied material.