visit-website-advanced
Public
Forked from vadimfedenko/visit-website-reworked
README.md
An advanced website-reading plugin for LM Studio with improved metadata extraction, structured headings, JSON-LD support, stronger reliability on modern websites, better diagnostics, and graceful fallback behavior.
This project is a fork of Visit Website Reworked.
The previous implementation worked on simple pages, but had issues on real-world websites:
<head> parsing could fail when the <head> tag had attributes.title, meta description, and headings could be empty even when present in the HTML.403, 401, 429, 503) had no recovery path.data-src, data-original, and srcset.r.jina.ai could contain useful markdown image links, but those were not always extracted as images.alt, source page URL, original image URL, local image path, local thumbnail path, and download time.Improved <head> parsing when the tag has attributes.
Improved title extraction.
Added meta description extraction from:
descriptionog:descriptiontwitter:descriptionAdded JSON-LD extraction from application/ld+json script tags.
Keeps JSON-LD as structured objects when parsing succeeds.
Falls back gracefully when invalid JSON-LD is encountered.
JSON-LD is extracted when available in the fetched HTML. It may be unavailable when fallback content is returned by r.jina.ai.
headingTree so H1/H2/H3 relationships are easier to understand.h1, h2, and h3 fields where possible.Added retry logic for direct page fetches.
Added automatic fallback via r.jina.ai when direct fetch is blocked or fails.
Added block/challenge detection for common anti-bot scenarios.
Image extraction behavior is inherited from the Reworked version and supports:
src, data-src, data-original, srcsetsource: "jina" casesVisit Website Advanced returns fetch metadata:
source (direct or jina)finalUrlstatusCodeserverErrors include more actionable status and server details.
websiteURL.Depending on the page and configuration, the tool can return:
urltitledescriptionh1h2h3headingTreejsonLdfetchlinksimagescompactGalleryMarkdownhintcontent1-click installation from LM Studio Hub:
yuanjumay/visit-website-advancedJust click Run in LM Studio.
You can tune:
Use 0 to exclude a section, or -1 for auto defaults where supported. Don't forget to allow the LLM to call the tool.
Ask the assistant to open a URL and extract page information.
For blocked pages, the plugin attempts fallback automatically and reports the fetch source.
Example prompts:
The LLM can use DuckDuckGo Advanced to find relevant URLs, then Visit Website Advanced to extract structured page content and images. It can then inspect image content via Analyze Images Advanced.
This project is a fork of Visit Website Reworked.
Original author:
Reworked version:
Advanced fork changes:
MIT
The original copyright and license notices are preserved.