Scrawly MCP Server

io.github.IliasSamiv1.0.0更新於 Oct 8, 2026

MCP server for the Scrawly SEO crawler: query local audits and propose confirmation-gated fixes.

概覽

AI 產生的概覽

讓助理查詢本機的 Scrawly SEO 爬取稽核結果,並針對網站技術問題提出需確認後才執行的修復建議。

功能
把 AI 助理連接到 Scrawly SEO 爬蟲,以便處理在本機執行的稽核。Scrawly 會爬取網站並執行 250 項技術 SEO、GEO 與代理式網路檢查,涵蓋可爬取性、重新導向、標題與中繼資訊、canonical、結構化資料、內部連結、效能以及 AI 搜尋就緒度。透過此 MCP 伺服器,助理可以查詢這些結果並提出修復方案,修復需經使用者確認後才會執行。
適用情境
適合已經執行 Scrawly 稽核、希望助理讀取結果、說明受影響頁面或草擬修復方案的情境。最適合自有或客戶網站的技術 SEO 工作,包括 AI 搜尋與 GEO 就緒度評估。如果只想做一次爬取而不需要助理參與,則不必安裝。
執行需求
以本機程序方式在使用者電腦上透過 stdio 執行,來自 PyPI 套件 sentinelseo;僅限桌面端,沒有網頁版可執行檔。Scrawly 桌面應用程式手動安裝需要 Python 3.11 或更新版本以及 git,並會一次性下載約 150 MB 的爬蟲瀏覽器。此 MCP 伺服器未宣告帳號、API 金鑰、環境變數或標頭。
安裝前請注意
修復屬於寫入動作:在 WordPress 網站上連接器可以修改標題、中繼描述、canonical 與重新導向,因此核准前應依賴確認提示與回復快照。稽核與報告只存放在本機,但應用程式會連線到你選擇稽核的網站、GitHub(檢查更新)以及你自行設定的服務,例如 PageSpeed Insights、Google Search Console 或 AI 供應商。此 MCP 伺服器本身未宣告任何憑證。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Scrawly MCP Server,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Desktop only,透過 STDIO。 STDIO 服務會啟動本機處理程序,因此需要 SourceWeft 桌面主機。

其他 MCP 客戶端

參照 儲存庫 中的啟動說明。

README

[Scrawly logo]

Scrawly: the free, open-source SEO crawler and site audit tool

Free technical SEO, GEO and AI-search audits for any website.
A free alternative to Screaming Frog and Sitebulb. No account, no page cap, no subscription.

[MIT License] [macOS, Windows and Linux] [250 SEO checks] [Latest release]

Scrawly is a free, open-source desktop SEO crawler and technical site audit tool for macOS, Windows and Linux. It crawls a website, runs 250 technical SEO, GEO (Generative Engine Optimization) and agentic-web checks on every page, maps the site's architecture, and explains in plain language what to fix. For WordPress sites it can also apply safe, reversible fixes. Scrawly is created by Ilias Sami and released for the SEO community under the MIT License.

In one sentence: Scrawly is a free desktop app that audits any website for the technical issues that affect Google rankings and AI-search visibility (ChatGPT, Perplexity, Google AI Overviews), with no account and no page limit.


Contents

Install in one line

Scrawly installs itself, adds an app icon (Applications on macOS, the Start Menu on Windows, your app menu on Linux), and keeps itself up to date.

macOS or Linux (Terminal):

bash
curl -fsSL https://raw.githubusercontent.com/IliasSami/scrawly-seo-crawler/main/install.sh | bash

Windows (PowerShell):

powershell
irm https://raw.githubusercontent.com/IliasSami/scrawly-seo-crawler/main/install.ps1 | iex

Then open Scrawly like any other app. There is nothing to sign up for.

Prefer to install by hand?

You need Python 3.11 or newer and git. (Node.js is not needed: releases ship the interface prebuilt.)

bash
git clone https://github.com/IliasSami/scrawly-seo-crawler.git Scrawlycd Scrawlypython3 scripts/install.py        # Windows: py scripts\install.py

The installer creates a private Python environment, downloads the crawler browser (about 150 MB, once), and adds Scrawly to your apps. Full guide, Linux notes and troubleshooting: docs/DESKTOP.md.

What Scrawly checks

Scrawly runs 250 checks grouped into 23 areas. Every finding comes with why it matters, how to fix it, and the exact pages affected.

AreaExamples
Crawlability and indexabilityrobots.txt rules, meta robots and X-Robots-Tag, noindex conflicts, orphan pages, crawl depth
Response codes and redirects4xx and 5xx pages, redirect chains and loops, temporary vs permanent redirects
Titles, meta and headingsmissing, duplicate, too long or too short titles; pixel-width truncation; meta descriptions too thin for AI; H1 and heading order
Canonicalizationmissing, conflicting, cross-domain or non-indexable canonicals
Content qualitythin and duplicate content, near-duplicates, readability
Structured dataJSON-LD, Microdata and RDFa, with schema validation
Internal linking and architecturelink graph, internal PageRank ("authority"), broken links, generic anchors
Performance and Core Web VitalsPageSpeed Insights lab and field (CrUX) data, TTFB, page weight
AI search readiness (GEO)which AI crawlers you allow or block, llms.txt, AI citation readiness
Agentic webhow well AI agents can read and operate your site
And moreimages and alt text, hreflang, XML sitemaps, security and TLS, mobile, JavaScript rendering, pagination, WordPress specifics

The full catalogue, with severity and fix tier for every check, is in docs/CHECK_MANIFEST.yaml.

Also included: JavaScript rendering, custom extraction, saved crawl profiles, a 3D site graph and depth tree, crawl-to-crawl comparison, scheduled audits with regression alerts, Google Search Console and GA4 data, branded HTML and PDF reports, an AI writer for titles and meta descriptions (bring your own AI key), and an MCP server so AI assistants can work with your audits (docs/MCP.md).

A free alternative to Screaming Frog and Sitebulb

Screaming Frog SEO Spider and Sitebulb are excellent tools. Scrawly is a free, open-source option for people who want a deep technical audit without a licence, plus checks for AI search.

Screaming Frog / SitebulbScrawly
PricePaid licence or subscription (limited free tier or trial)Free and open-source (MIT)
Account or licence keyNeeded for full useNone
Page limitThe free tier is cappedNo cap: you set the limit
AI search (GEO) and agentic-web checksSome, depending on versionA dedicated set: AI crawler access, llms.txt, citation and agent readiness
FixesReport onlySafe, reversible fixes on WordPress
Source codeClosedOpen: read, audit, contribute

Scrawly is independent and not affiliated with Screaming Frog Ltd or Sitebulb.

GEO and AEO: auditing for AI search

Search now includes AI answer engines that read your pages and decide whether to cite you. Scrawly audits for Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO):

  • AI crawler access: a matrix of which AI bots (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot and others) your robots.txt and firewall allow or block.
  • llms.txt: whether you publish a curated index for AI systems, and its quality.
  • Meta as AI context: a description is often the first thing an AI reads about a page, so Scrawly flags descriptions too thin to understand and can draft an entity-rich one from the page's own content.
  • Entities and structured data: the schema and named entities that help search engines and AI understand what a page is about.

Fix WordPress issues safely

Connect a WordPress site with the lightweight Scrawly Connector plugin and Scrawly can fix titles, meta descriptions, canonicals and redirects for you:

  • a rollback snapshot is saved before every change, and every fix can be undone;
  • changes to the live site always ask for your confirmation first;
  • higher-risk issues are never auto-applied; they're flagged for a person to review.

See docs/WORDPRESS_CONNECTOR.md. Every other site (Shopify, Wix, Webflow, Next.js, custom) gets the full read-only audit.

Privacy: your audits stay on your computer

Scrawly has no accounts and no analytics. Your crawls, findings and reports are stored on your computer only. Scrawly connects to:

  • the websites you choose to audit;
  • GitHub, to check for updates (turn off with SCRAWLY_AUTO_UPDATE=0);
  • services you configure, such as PageSpeed Insights, Google Search Console or your AI provider;
  • the feedback service, only when you click Send on a feedback note.

The app's local engine only accepts requests from the Scrawly window itself, so other websites can't use it. Report security issues as described in SECURITY.md.

Frequently asked questions

What is Scrawly? Scrawly is a free, open-source SEO crawler and technical site audit tool that runs as a desktop app on macOS, Windows and Linux. It checks every page of a website for 250 technical SEO, GEO and agentic-web issues and explains how to fix them.

Is there a free alternative to Screaming Frog? Yes. Scrawly is a free, open-source alternative to Screaming Frog SEO Spider and Sitebulb. It has no page cap and no licence, and it adds AI-search (GEO) checks and safe WordPress fixes.

Is Scrawly really free? Yes. Scrawly is free and open-source under the MIT License, with no account, subscription or usage limit. Audits run on your own computer, so there is no per-crawl cost.

Do I need to create an account? No. Install it and start auditing. There is no sign-up, login or approval.

Does Scrawly work on websites that aren't WordPress? Yes. The audit works on any website. Only the one-click fixes are WordPress-specific.

Can Scrawly check AI search and GEO readiness? Yes. It checks AI crawler access, llms.txt, whether titles and descriptions give AI enough context, structured data, and how well AI agents can use the site.

How many pages can Scrawly crawl? There is no fixed cap. You choose the limit per audit; how far you can go depends on your computer and the website.

Who made Scrawly? Scrawly was created by Ilias Sami, an SEO consultant, as a free contribution to the SEO community. Contact: [email protected].

Documentation

Contributing and feedback

Scrawly gets better from community input:

  • In the app: click the speech-bubble icon (Send feedback) at the bottom of the left bar. No GitHub account needed.
  • On GitHub: open an issue or a pull request. See CONTRIBUTING.md.

License and credits

Scrawly is released under the MIT License. Copyright (c) 2026 Ilias Sami ([email protected]). If you use Scrawly in research or writing, see CITATION.cff.

Built for the SEO community. If Scrawly helps you, a star on GitHub helps others find it.

來源:README.md,提交 123b766

工具

0
工具後設資料尚未被收錄。

版本歷史

1
  1. v1.0.0最新Oct 8, 2026