crawlora-mcp

# Crawlora MCP

[![Website](https://img.shields.io/badge/website-crawlora.net-2563eb)](https://crawlora.net/mcp?utm_source=github&utm_medium=referral&utm_campaign=crawlora-mcp)
[![Docs](https://img.shields.io/badge/docs-crawlora.net%2Fdocs-555)](https://crawlora.net/docs?utm_source=github&utm_medium=referral&utm_campaign=crawlora-mcp)
[![MCP](https://img.shields.io/badge/MCP-Streamable%20HTTP-7c3aed)](https://modelcontextprotocol.io)
[![smithery badge](https://smithery.ai/badge/crawlora/crawlora)](https://smithery.ai/servers/crawlora/crawlora)
[![crawlora-mcp MCP server](https://glama.ai/mcp/servers/Crawlora-org/crawlora-mcp/badges/score.svg)](https://glama.ai/mcp/servers/Crawlora-org/crawlora-mcp)

**Crawlora MCP** is a **hosted** Model Context Protocol server that gives AI clients and
agents **1873 structured public‑web‑data tools** across 215 platform groups — search, maps,
e‑commerce, social, finance, travel, app stores, media, and reviews — each returning clean,
normalized **JSON** instead of HTML to parse.

> **Two ways to use it:** connect any MCP client to the **hosted** endpoint
> (`https://mcp.crawlora.net/mcp`), or run the small **open‑source local server** in this repo
> (`npx` / Docker, stdio) — both expose the same tools and authenticate with your Crawlora API
> key. Start free with **2,000 credits/month** (no card) at **[crawlora.net](https://crawlora.net?utm_source=github&utm_medium=referral&utm_campaign=crawlora-mcp)**.

## Connection

| | |
|---|---|
| **Endpoint** | `https://mcp.crawlora.net/mcp` |
| **Transport** | Streamable HTTP |
| **Auth** | `Authorization: Bearer <CRAWLORA_API_KEY>` (preferred) — `x-api-key: <key>` also accepted |
| **Get a key** | [https://crawlora.net](https://crawlora.net?utm_source=github&utm_medium=referral&utm_campaign=crawlora-mcp) (free 2,000 credits/mo) |
| **Server card** | https://crawlora.net/.well-known/mcp/server-card.json |

A missing or invalid API key returns `401`.

## Connect your client

Crawlora MCP works with any MCP‑capable client. Pick yours below — they all point at the same
hosted endpoint (`https://mcp.crawlora.net/mcp`, Streamable HTTP) and authenticate with your
`CRAWLORA_API_KEY`. Prefer keeping the key in an environment variable over pasting it literally,
and never commit it. Ready‑to‑paste files for each client live in [`examples/`](examples/).

### Claude Code (CLI)

```bash
claude mcp add --transport http crawlora https://mcp.crawlora.net/mcp \
  --header "Authorization: Bearer ${CRAWLORA_API_KEY}"
```

Adds at `local` scope by default. Use `--scope user` to make it available in every project, or
`--scope project` to write a shared `.mcp.json`. Confirm with `claude mcp list`.

### Claude Desktop

Settings → Developer → **Edit Config**, then add to `claude_desktop_config.json`:

```json
{
  "mcpServers": {
    "crawlora": {
      "url": "https://mcp.crawlora.net/mcp",
      "transport": "streamable-http",
      "headers": { "Authorization": "Bearer YOUR_CRAWLORA_API_KEY" }
    }
  }
}
```

### Cursor

Project‑local `.cursor/mcp.json` (or `~/.cursor/mcp.json` for all projects):

```json
{
  "mcpServers": {
    "crawlora": {
      "url": "https://mcp.crawlora.net/mcp",
      "headers": { "Authorization": "Bearer ${CRAWLORA_API_KEY}" }
    }
  }
}
```

### Codex CLI

```bash
codex mcp add crawlora \
  --url https://mcp.crawlora.net/mcp \
  --bearer-token-env-var CRAWLORA_API_KEY
```

### VS Code (GitHub Copilot)

`.vscode/mcp.json` (project) or your user `mcp.json`:

```json
{
  "servers": {
    "crawlora": {
      "type": "http",
      "url": "https://mcp.crawlora.net/mcp",
      "headers": { "Authorization": "Bearer ${env:CRAWLORA_API_KEY}" }
    }
  }
}
```

### Windsurf

`~/.codeium/windsurf/mcp_config.json` (note the `serverUrl` key):

```json
{
  "mcpServers": {
    "crawlora": {
      "serverUrl": "https://mcp.crawlora.net/mcp",
      "headers": { "Authorization": "Bearer ${env:CRAWLORA_API_KEY}" }
    }
  }
}
```

### Gemini CLI

`~/.gemini/settings.json` (note the `httpUrl` key):

```json
{
  "mcpServers": {
    "crawlora": {
      "httpUrl": "https://mcp.crawlora.net/mcp",
      "headers": { "Authorization": "Bearer YOUR_CRAWLORA_API_KEY" }
    }
  }
}
```

### Cline / Roo Code

Add to `cline_mcp_settings.json` using the generic remote shape — full walkthrough in
[`llms-install.md`](llms-install.md):

```json
{
  "mcpServers": {
    "crawlora": {
      "url": "https://mcp.crawlora.net/mcp",
      "transport": "streamable-http",
      "headers": { "Authorization": "Bearer ${CRAWLORA_API_KEY}" }
    }
  }
}
```

### Other clients

Any client that speaks **Streamable HTTP** can use the generic config in
[`examples/mcp.json`](examples/mcp.json). **stdio‑only** clients (e.g. Zed) should run the
[local server](#run-it-locally-open-source-server) below and point at its `npx` command.

## Run it locally (open‑source server)

This repo also ships a small **stdio** MCP server (`index.mjs`) that exposes the same **1873
tools**, each wrapping the Crawlora REST API (`https://api.crawlora.net/api/v1`) with your
`CRAWLORA_API_KEY`. Useful if you'd rather run the server yourself than use the hosted endpoint.

```bash
# Node 20+ (no install)
CRAWLORA_API_KEY=your-key npx -y crawlora-mcp

# or from a clone
npm install && CRAWLORA_API_KEY=your-key node index.mjs

# or Docker
docker build -t crawlora-mcp . && docker run -i -e CRAWLORA_API_KEY=your-key crawlora-mcp
```

Client config (stdio):

```json
{
  "mcpServers": {
    "crawlora": {
      "command": "npx",
      "args": ["-y", "crawlora-mcp"],
      "env": { "CRAWLORA_API_KEY": "your-key" }
    }
  }
}
```

The tool definitions in [`tools.json`](tools.json) are generated from Crawlora's published API
catalog; each carries its real input schema and maps to a single REST endpoint.

## Hosted MCP (recommended)

Just paste the URL into your client settings. No Node.js required.

```text
https://mcp.crawlora.net/mcp
```

Header:
`x-api-key`: `<your-api-key>`

Get a key at [crawlora.net](https://crawlora.net?utm_source=github&utm_medium=referral&utm_campaign=crawlora-mcp).
Every account includes **2,000 free credits / month** with no credit card required.

---

## Tools

The complete tool catalog is generated from [`tools.json`](./tools.json) so MCP directories can index it.

<!-- BEGIN MCP.SO TOOLS: generated from tools.json; do not edit manually -->
### `accor_amenities`

Accor amenity reference catalog. Returns the anonymous public Accor amenity catalog with stable amenity codes, labels, categories, and display ordering. It is reference data for interpreting hotel search facets; booking, rates, rooms, reviews, and account data are excluded.

### `accor_brands`

Accor brand directory. Returns the Accor brand directory published on the public brands page, with each brand's name, slug, and source URL. Booking, rates, rooms, reviews, contacts, and location are excluded.

### `accor_catalog_hotels`

Search the Accor hotel catalog. Searches Accor's anonymous public hotel catalog by text, hotel code, or latitude/longitude radius. Results contain static property identity and broad location metadata plus catalog relevance/distance; contact details, precise hotel coordinates, payment, loyalty, media, rates, rooms, reviews, and booking data are excluded.

### `accor_destination_hotels`

Accor hotels in a destination. Returns the static list of hotels published on one Accor destination directory page (world, continent, country, region, department, city, district, or place), optionally narrowed by a theme facet. Each hotel carries its code, name, source URL, and broad city/country. Booking, rates, rooms, reviews, contacts, and precise location are excluded.

### `accor_property`

Accor hotel metadata. Returns static public metadata for one Accor hotel code, including name, brand, city/country, explicitly listed amenities, and published check-in/check-out times. Booking, rates, rooms, reviews, contacts, and precise location are excluded.

### `accor_search`

Search Accor hotels. Searches the public Accor hotel index by destination or hotel name, with optional country, city, brand, star, page, and page-size filters. Results contain static hotel identity, location labels, ratings, and source URLs; booking, rates, rooms, reviews, loyalty, and payment flows are excluded.

### `accor_search_details`

Accor place coordinates and viewport. Resolves an anonymous Accor search suggestion identifier to its description, coordinates, viewport, radius, and address components. It uses the same public place-details source as the Accor search box; booking, rates, rooms, reviews, and account data are excluded.

### `accor_search_suggest`

Accor destination and hotel search suggestions. Returns anonymous Accor typeahead suggestions for a partial destination or hotel query, including destination/place and hotel identifiers, types, labels, and match metadata. It uses the public search-box suggestion source only; booking, rates, rooms, reviews, and result-list requests are excluded.

### `adidas_product`

Get an Adidas product. Returns normalized product-detail data for one Adidas SKU: name, brand, category, description, pricing (current/standard/sale), images, and every purchasable size variant. product_id is the Adidas SKU (e.g. JI0397), taken from a search result's products[].id field or the trailing segment of an Adidas product page URL. An unknown product_id returns a not-found error.

### `adidas_product_review_topics`

Get Adidas review topics for a product. Returns the topics an Adidas product model's customer reviews can be filtered by -- the "filter by topic" chips the product page shows, commonly satisfaction, comfort, color, purchase, fit, appearance, quality and style. Feed a topics[].topic value back to /adidas/product/reviews as its topic parameter to return only reviews about that aspect. The topic vocabulary is per model, not a fixed list: a shoe exposes fit and comfort topics that an accessory does not, so read it