Crawleo Web Search — Crawleo 网页 搜索

Crawleo OpenClaw 技能

Use this 技能 when OpenClaw needs Crawleo-powered live 网页 搜索, Google 搜索 SERP data, Google Maps place data, URL crawling/content 提取ion, or headful browser crawling for 保护ed and highly dynamic sites.

Current Implementation 状态

This 仓库 now includes offline-tested Crawleo REST wr应用er 辅助工具s for all five documented 端点s. Live Crawleo calls require CRAWLEO_API_KEY; user-facing examples and optional live verification are expanded in later slices. Do not clAIm live behavior was verified unless an explicitly enabled live test has 运行.

Source of Truth

Use contracts/crawleo-端点s.json as the machine-readable 端点 contract, contracts/crawleo-端点s.md as the human-readable contract, and contracts/coverage-检查列出.md as the 端点-to-wr应用er/test/example coverage 检查列出. These files cover:

/搜索 m应用ed to MCP 工具 搜索_网页 /google-搜索 m应用ed to MCP 工具 google_搜索 /google-maps m应用ed to MCP 工具 google_maps /crawl m应用ed to MCP 工具 crawl_网页 /headful-browser m应用ed to MCP 工具 headful_browser

Crawleo 端点-specific docs take precedence over the local OpenAPI snapshot when sources conflict. If a default, limit, 响应 field, error table, or parameter is unclear, write not specified in Crawleo docs rather than inventing behavior.

Authentication and Secret Handling

Live Crawleo REST calls require CRAWLEO_API_KEY. 发送 it with Crawleo's documented x-API-key header by default. Crawleo also documents Authorization: Bearer YOUR_API_KEY as an alternate authentication style.

创建 a 命令行工具ent with 创建Crawleo命令行工具ent({ APIKey: process.env.CRAWLEO_API_KEY }), then call 命令行工具ent.搜索, 命令行工具ent.google搜索, 命令行工具ent.googleMaps, 命令行工具ent.crawl, or 命令行工具ent.headfulBrowser with 端点 parameters from contracts/crawleo-端点s.json.

Never print, echo, 记录, persist, or include API key values in errors, examples, test 输出, or 调试 输出.

Offline-First Behavior

Default commands, examples, and tests must be offline-safe. They must not call https://API.crawleo.dev, require CRAWLEO_API_KEY, or consume Crawleo credits unless explicitly marked as live tests.

Live tests must require 机器人h:

CRAWLEO_API_KEY CRAWLEO_ENABLE_LIVE_TESTS=1

运行 optional live verification with:

CRAWLEO_API_KEY=... CRAWLEO_ENABLE_LIVE_TESTS=1 npm 运行 test:live

Without 机器人h variables, the live test skips safely and exits 0.

端点 Use 图形界面dance Use /搜索 / 搜索_网页 for Bing-powered 网页 搜索 with optional auto-crawling and content 提取ion for LLM/RAG 工作流s. Use /google-搜索 / google_搜索 for Google SERP data, including 网页, news, images, places, shopping, knowledge graph, People Also Ask, related 搜索es, and answer boxes. Use /google-maps / google_maps for structured place/business/landmark data from Google Maps. Use /crawl / crawl_网页 for direct URL crawling and content 提取ion. Try this before headful browser to reduce credit usage. Use /headful-browser / headful_browser only when standard crawling is blocked or a headed browser/screenshot path is required. Crawleo docs say this costs 50 credits per URL and fAIled 请求s cost 0 credits. Verification

运行 offline verification with:

npm test npm 运行 验证:contracts npm 运行 验证:examples npm 运行 验证:scaffold

At this stage, npm 运行 验证:scaffold proves the self-contAIned package files exist, point to the Crawleo contract inventory, and 导出 the 运行time wr应用er surface. Later slices 添加 richer examples, documentation, and optional live-test gating.