Community publisher · 11 products
Products
Markdown Cleaner for RAGClean Markdown documents to boost retrieval quality in Dify knowledge bases (parent-child chunking). Supports both file upload (.md) and direct text input from previous workflow nodes. A 12-step pipeline removes header/footer noise, fixes cross-page sentence breaks, protects clause numbers, repairs garbled symbols, normalizes terminology, and deduplicates chunks by MD5.tool · v0.0.4Table Format ConverterConvert tables across Excel/CSV, JSON, and Markdown in four directions (Excel/CSV→JSON, Excel/CSV→Markdown, JSON→Excel/CSV, Markdown→Excel/CSV), plus Excel/CSV to Array (row-object array for Iteration nodes), with sheet, output-format, and row-limit options.tool · v0.0.12ECharts Shine ChartsRender publication-ready ECharts visualizations in the elegant Shine theme. Covers 30+ chart types — bar, line, pie, scatter, radar, heatmap, funnel, gauge, candlestick, sankey, tree/treemap, sunburst, parallel, boxplot, graph, themeriver, calendar, and 3D/GL charts — plus built-in China national & provincial geo-maps. All ECharts assets are embedded for fully offline rendering.tool · v0.0.3Knowledge Graph BuilderTurn Markdown documents into knowledge-graph-style chunks for Dify knowledge bases: each chunk carries entity and relation annotations to boost multi-hop LLM reasoning and retrieval recall. Includes a validator that checks entity coverage, relation completeness, and trap blocks, and outputs a quality report.tool · v0.0.3Excel/CSV Splitter for LLMSplit large Excel/CSV files into LLM-friendly parts to avoid context-window overflow. Supports four modes — by size (MB per part), by count (N equal parts), by rows (every N rows), and by sheet — with CSV or Excel output. Designed to feed Dify iteration nodes part by part.tool · v0.0.6PyMuPDF PDF to Markdown / JSONConvert digital-native PDFs into clean Markdown or structured JSON locally with PyMuPDF — detects headings, tables, and reading order, strips repeated headers/footers, and supports page-range selection. Pure Python, no model files, runs on amd64 and arm64.tool · v0.0.2Markdown to ReportConverts Markdown text into a professional Word (.docx) or single-file HTML report with sober-ink design style, suitable for executive reporting.tool · v0.0.1HTTP Big-Data Slimmer for LLMSlim, filter and paginate large HTTP responses before sending to LLM.tool · v0.0.4HTML Preview RendererWraps any HTML string into a self-contained page for preview. Accepts raw HTML (e.g. ECharts output) and returns it as a renderable HTML page.tool · v0.0.2Lightweight Cache (Redis-style)An in-process key-value cache with Redis-style SET / GET / DEL / EXISTS and TTL expiry — no real Redis server required, ideal for air-gapped intranets. Data is persisted to disk and survives container restarts.tool · v0.0.2Recursive Web CrawlerRecursively crawl web pages from a starting URL and depth, returning structured JSON for each page (title, main content as Markdown, links, and metadata). Supports same-domain restriction, max-page limits, and URL include patterns; tolerant of self-signed intranet certificates.tool · v0.0.2