DataPluck: Web Scraper & JSONL Dataset Builder

Point, click, export. Turn any list, feed, table or article into clean CSV, Markdown, JSON or JSONL, right in your browser.

Turn any webpage list, table, feed, Q&A thread, or article into clean, structured data in one click. DataPluck is a visual web scraper and dataset builder that runs 100% locally in your browser with no external backend. HOW IT WORKS 1. Open any page with a list, table, feed, or article and click the DataPluck icon (or press Alt+Shift+P). 2. Hover to preview repeating items and click once to lock your selection. 3. Clean, rename, or toggle columns, then click Download or Copy. CORE FEATURES (FREE & PRO) • Smart Visual Picker: Automatically detects repeating sibling items, tables, and items spread across multiple sections. Step between Broader and Narrower levels with a click or the Up/Down arrow keys. • Auto Columns: Locates fields by structural position, names them intelligently (price, date, rating, headings), and automatically extracts link and image columns while switching off constant boilerplate columns (like "Add to cart"). • Per-Column Data Cleaning: Parse currencies and numbers ($1,299.00 and 1.299,00 €), convert to lower/upper case, extract link domains, find and replace text, or extract values using regular expressions with a live before-and-after row preview. • Deduplication & Provenance: Count and remove duplicate rows on the fly. Append source_url and captured_at columns for dataset auditing. • Article to Clean Markdown: Strip scripts, styles, hidden nodes, navbars, footers, ads, cookie banners, share bars, and tracking parameters (utm_*, fbclid, gclid) into clean Markdown with a live token estimate. • Spreadsheet Formula Safety: Automatically prefixes cells starting with =, +, -, or @ so scraped pages cannot plant malicious formulas into your CSV or Google Sheets exports. • Crash Recovery: Long multi-page runs are mirrored to local browser storage. If a tab or browser closes before you export, DataPluck offers to restore your run on your next visit. PRO AUTOMATION & AI FORMATS • Unlimited Rows: Remove the 50-row free export cap for large-scale data collection. • Multi-Page Pagination (Up to 200 Pages): Automatically follows "Next" buttons or numbered pagers (1, 2, 3...) across single-page apps (SPAs) and full page loads. Matches shifting rows by their unique link to prevent duplicates across pages. • Infinite Scroll & Virtualized Lists: Captures rows after every small scroll step so virtualized feeds that remove off-screen elements from the DOM are collected completely. • LLM Fine-Tuning & RAG Formats: Export JSON, JSONL, RAG Markdown chunks with token counts, or Q&A threads formatted for OpenAI chat messages, prompt/completion, Alpaca (instruction, input, output), and ShareGPT conversations. • Google Sheets & Notion Export: Copy tables as TSV and open sheets.new in one click (no OAuth required), or send data directly to a Notion sub-page using your Notion integration token. PRIVACY FIRST Everything runs locally inside your browser. No page content ever leaves your computer unless you explicitly export it (such as to your own Notion workspace).