使用方法
- 复制安装命令，在自己的智能体环境安装前先检查源码包。
- 从原始文档核对所需工具、凭据和支持的宿主。
- 说明输入材料与期望交付物，并对照来源检查输出。
npx skills add https://github.com/fetcher-sh/fetcher-skills --skill google-search开始前需了解
- 安装不包含模型订阅、API 额度或访问文件的权限。
- 仓库星数和安装量反映活跃情况，不等于独立质量或安全测试。
按源仓库内容呈现，保留标题、案例、代码、表格、链接以及原文引用的演示图片。
Google Search API
Google's own search results as clean JSON — one plain HTTP GET per call, paid as you go. Google's search operators pass straight through: site:, filetype:, intitle:, and quoted exact phrases all work exactly as they do in the browser. No Custom Search Engine setup, no per-day query quota, no billing account.
Base URL: https://google.fetcher.sh
Authentication
Two ways to pay, same data — full mechanics in the `fetcher` skill:
# 1. Prepaid credits (recommended — get a key at https://fetcher.sh/topup
# or via POST /api/credits/topup, see the fetcher skill)
export FETCHER_API_KEY="bby_live_xxxxxxxxxxxx"
curl -H "Authorization: Bearer $FETCHER_API_KEY" \
"https://google.fetcher.sh/api/search?query=hello"
# 2. x402 pay-per-call — omit the header; a GET with no payment returns 402
# with machine-readable payment requirements (USDC on Base, Polygon,
# Arbitrum, Monad, or Solana). @x402/fetch signs and retries automatically.Every response is { "status": number, "message": string, "data": ... }; the HTTP status mirrors status.
Endpoint ($0.005/call)
| Endpoint | What it returns |
|---|---|
/api/search | Google web search results for a query |
Required: query. Optional: safe (SafeSearch), page (pagination), hl (UI language, e.g. en), country (region scope), count (results per page — currently 10), noEncode (skip query re-encoding for pre-formatted operator strings).
Scenarios
Restrict to one site:
curl -H "Authorization: Bearer $FETCHER_API_KEY" \
--data-urlencode "query=site:github.com x402 payments" -G \
"https://google.fetcher.sh/api/search"PDFs only:
curl -H "Authorization: Bearer $FETCHER_API_KEY" \
--data-urlencode "query=agentic commerce filetype:pdf" -G \
"https://google.fetcher.sh/api/search"Exact phrase:
curl -H "Authorization: Bearer $FETCHER_API_KEY" \
--data-urlencode 'query="pay per call api"' -G \
"https://google.fetcher.sh/api/search"Localized results (UK, French UI) and pagination:
curl -H "Authorization: Bearer $FETCHER_API_KEY" \
--data-urlencode "query=agentic commerce" -G \
--data-urlencode "country=GB" \
--data-urlencode "hl=fr" \
--data-urlencode "page=2" \
"https://google.fetcher.sh/api/search"Combine operators — title match plus site restriction:
curl -H "Authorization: Bearer $FETCHER_API_KEY" \
--data-urlencode "query=intitle:pricing site:x402.org" -G \
"https://google.fetcher.sh/api/search"MCP
{
"mcpServers": {
"google-search": {
"url": "https://google.fetcher.sh/mcp",
"headers": { "Authorization": "Bearer bby_live_..." }
}
}
}Free: search_endpoints, describe_endpoint, check_balance. Paid: fetch_data (any endpoint above), topup_credits, plus the named shortcut google_search. Drop the headers block to pay per call with x402 instead — see the `fetcher` skill for the full flow.
Errors
400— missing/invalid parameter (message names it)401— unknown or rotated key402— payment required (x402 challenge) ortopup_required(credits
exhausted)
404— not a priced path- No rate limits; no refunds on upstream failures (settlement precedes
delivery)
Reference
- Full agent setup: <https://google.fetcher.sh/skill.md>
- OpenAPI 3.1 contract: <https://google.fetcher.sh/openapi.json>
- Condensed catalog: <https://google.fetcher.sh/llms.txt>
- Payment, credits, and MCP deep dive: `fetcher` skill
- Site: <https://google.fetcher.sh>
