[DibbayajyotiRoy]

Why

People find Dify plugins by searching the Marketplace and the web for what they want to do ("web scraper", "URL to Markdown", "scrape JavaScript sites", "extract data"). The listing led with the brand name and described the tools in our own terms. This reworks the copy the Marketplace renders so it matches those searches, without claiming anything the plugin does not do.

Where each field renders (checked against the live listing at marketplace.dify.ai/plugins/spicrawl/spicrawl)

Field	Shown as
label	Page <h1>, <title> ("... - Dify Marketplace"), card name, tool picker in Dify
description	Card summary, <meta name="description">, og:description
tags	Category filters
README.md	The body of the detail page
Tool label / human description	Tool picker and node in Dify
Tool llm description	Read by agents when they choose a tool

Before / after

Label

• Before: Spicrawl
• After: Spicrawl Web Scraper (also zh_Hans, ja_JP)

Description (160 characters, lead keywords first)

• Before: "Scrape any URL to clean markdown, text or HTML, pull fields out of a page with CSS selectors, and run batch scrapes with the Spicrawl web scraping API."
• After: "Web scraper for Dify. Scrape any website to clean Markdown for LLMs, RAG and AI agents, render JavaScript pages, extract fields and batch scrape up to 100 URLs." (also zh_Hans, ja_JP)

Tags (all from the daemon's PluginTag list; the provider tags use the SDK's narrower ToolLabelEnum, and all four are in both)

• Before: search, utilities
• After: search, utilities, rag, business
• Not used because they would be untrue: agent and trigger (those are agent-strategy and trigger plugins), other, and the content categories (image, videos, weather, ...). productivity is left out because nothing in the docs backs it.

Tools

Before	After
Scrape	Scrape Web Page
Extract	Extract Page Data
Batch scrape	Batch Scrape URLs

Each has a keyword-led human description and a fuller llm description (with zh_Hans and ja_JP for label and human description). Parameters are unchanged.

README

An answer-first opening ("To scrape a website in Dify, install the Spicrawl plugin, add a free API key and use Scrape Web Page ..."), then What you can do, a 3-step quick start, a compact tools table, a 5-question FAQ (JavaScript sites, RAG, cost, field extraction, crawling), errors and security, and links. The setup, credentials, connection requirement, source and privacy content the review guidelines ask for is all still there.

Accuracy notes

• Cost claims match docs.spicrawl.com/credits and the pricing page: free plan 1,000 credits a month, no card; plain page 1 credit, JavaScript 3; failures, bot-challenge pages and error-status pages cost nothing. The README does not say empty pages are free (an empty 200 page is billed; the FAQ tells users to switch on JavaScript rendering instead).
• The plugin does not follow links, so the README says so and points to Batch Scrape URLs for many pages; "crawl" is not used as a feature claim.
• No engine or browser product names, no competitor names, no superlatives.

Version

0.1.2 -> 0.1.3 (manifest and User-Agent). The 0.1.2 icon is unchanged. No functional change: tool parameters and code paths are the same.

Checks

• pytest in dify/: 16 passed.
• Packaged with Dify CLI 0.6.11 and run through dify-marketplace-toolkit validate-difypkg.py: 0 blocking failures, 0 environment failures, the same 3 warning categories as 0.1.2 (all pre-existing).

After merge: tag dify-v0.1.3, then one marketplace update PR with spicrawl-0.1.3.difypkg.