Lookout

io.github.SylphxAI由 GitHub 账号 SylphxAI 发布v0.3.3更新于 Oct 24, 2026

Lookout — web answers with source-level proof. Search and fetch citeable excerpts, no API key.

已验证STDIO仅桌面网页搜索与抓取知识与记忆

概览

AI 生成的概览

Lookout 是一个本地 MCP 服务器，让助手无需 API 密钥即可进行网页搜索、抓取和提取，返回可引用的摘录，并在抓取失败时明确报告。

可以做什么
通过四个无需密钥的公开适配器（DuckDuckGo HTML、Wikipedia OpenSearch、npm registry、Hacker News Algolia）搜索网页，返回排序后的结果并标明所用适配器。抓取单个 URL，带 SSRF 检查、重定向与大小上限，返回正文前缀片段，并对被拦截地址、非 200 响应、截断和空提取给出警告或错误。从 HTML 或 URL 中提取标题、元数据、表格和片段。执行有边界的同源爬取，或运行命名的多步研究工具，先搜索再抓取并提取排名靠前的页面（默认 3 个，最多 6 个）。逐词比较两段文本或两个 URL，并查询或管理磁盘缓存。
适用场景
适合希望助手基于实际打开的页面作答、而不是转述搜索摘要的场景，也适合希望把失败如实报告为缺口而非给出自信总结的情况。适用于研究、事实核查和阅读公开网页，且无需注册搜索 API。
运行要求
通过 stdio 在本地运行，通常用 npx 从 npm 包启动，因此需要 Node.js 和网络访问。无需 API 密钥，未声明环境变量或请求头。首次搜索仍需联网；之后的搜索和抓取可能由磁盘缓存提供。
安装前请注意
抓取和爬取会向第三方网站发送请求，搜索查询会发送给四个公开适配器。抓取默认不遵守 robots.txt，除非设置 respectRobots；爬取则默认遵守，除非关闭。搜索和抓取默认启用缓存，命中缓存时会从磁盘重放此前的响应。README 说明该仓库已归档并合并到另一工具，使用前请确认维护状态。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Lookout，将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Desktop only，通过 STDIO。 STDIO 服务会启动本地进程，因此需要 SourceWeft 桌面宿主。

其他 MCP 客户端

参照 仓库 中的启动说明。

README

Merged into anymd (2026-09-25). anymd reads PDFs, Office files, EPUB, web pages, images and video into clean Markdown for AI agents: npx -y @sylphx/anymd. This repository is archived.

[Lookout]

Lookout

Web answers with source-level proof.

Search and fetch citeable excerpts. No API key. Hits name the adapter. Fetches keep a span, warnings, and gaps. Multi-step research is a separate tool.

[npm] [license]

npm @sylphx/lookout · bin lookout · MCP io.github.SylphxAI/lookout

The problem

A search snippet is a title and a URL. The agent then paraphrases a page it never opened. When the body is missing, a confident summary is worse than a gap.

The difference

A snippet saysLookout returns
“Release notes” and a URLtitle, description, excerpt span, and any table rows it could read
nothing when the fetch failsa warning or an error: blocked address, HTTP status, truncation, empty extract

Fetch reads the HTTP response. It does not run JavaScript, and it does not write a model summary.

Search, then open the page

bash
npx -y @sylphx/lookout search "Model Context Protocol"npx -y @sylphx/lookout fetch https://example.comnpx -y @sylphx/lookout extract https://example.com

search queries DuckDuckGo HTML, Wikipedia OpenSearch, the npm registry, and Hacker News Algolia, then ranks the hits. It does not fetch those pages and it does not run web_research.

Install

bash
npx -y @sylphx/lookout

That starts a stdio MCP server. No API key.

Your clientSetup
Any agent / CLInpx -y @sylphx/lookout
Claude Codeclaude mcp add lookout -- npx -y @sylphx/lookout
Claude Desktop / Cursor / VS Code / Codex"command": "npx", "args": ["-y", "@sylphx/lookout"]
json
{  "mcpServers": {    "lookout": { "command": "npx", "args": ["-y", "@sylphx/lookout"] }  }}

Tools

ToolWhen to call it
web_searchRanked hits from the four public adapters. No API key.
web_fetchOne URL. SSRF checks, redirect cap, size cap, body-prefix span.
web_extractTitle, metadata, tables, and spans from HTML or a URL.
web_cacheQuery or manage the on-disk cache. Not a web search.
web_crawlBounded same-origin crawl. Not the search default.
web_researchNamed multi-step tool: search, then fetch and extract the top pages (default 3, max 6).
web_diffAdded and removed words between two texts or two URLs.

There is no profile switch. The tool you name is the work that runs. Cache is on for search and fetch unless you pass useCache: false. A cache hit replays an earlier response from disk. The first search still needs the network.

Reference: tools · defaults

What it will not pretend

  • Private, loopback, and link-local addresses are rejected.
  • A failed adapter, a non-200 response, a truncated body, or an empty extract is reported. It is not turned into a clean answer.
  • web_fetch does not honor robots.txt unless you set respectRobots: true. web_crawl honors it unless you turn that off.
  • Lookout does not render a browser and does not synthesize an answer.

Companion MCP tools

ProductJob
CitraPDF answers with page-level proof
IrisImage facts and pixel evidence
CueVideo timelines and timestamp evidence
SpineRepository architecture and impact
LocusExact code-chunk retrieval

Each product is independent. Install only the tools your agent needs.

Documentation

Websitesylphxai.github.io/lookout
QuickstartInstall and first call
CompareFirecrawl, Tavily, and what Lookout actually does

Development

bash
bun installbun testbun run doctorbun run docs:build

License

MIT

来源：README.md，提交 87ae52c

工具

0
SourceWeft 尚未在沙箱中启动过该软件包，因此其工具未知。

版本历史

3
  1. v0.3.3最新Sep 23, 2026
  2. v0.3.2Sep 23, 2026
  3. v0.3.1Sep 23, 2026

评分

已满足适用检查项的 50%

该评分衡量此收录条目的完整程度，依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价，也不表示服务器安全或当前可用：可连通性是带日期的检查结果，评分也不包含安装量或用户评价。

  • 近 30 天内可连通25 分

    请保持端点能够响应 MCP 握手；SourceWeft 每 30 天探测一次。

  • 已声明认证方式10 分

    请在注册中心条目中声明凭据请求头，或保持端点能够响应探测，以便判定其认证方式。

  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list，或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分