Quick Data Scraper

Extract tables and lists from any page. Preview columns, paginate or scroll, export CSV/XLSX. All local in your browser.

Quick Data Scraper is a Chrome Manifest V3 extension that finds structured-ish data on the active web page—HTML tables and repeated sibling blocks—and lets the user preview, filter, rename or hide columns, optionally merge multiple “pages” (next link or infinite scroll), and download CSV or XLSX. Extraction runs in the page context via an injected content script; exports are built in the popup using bundled SheetJS (xlsx); files are offered through chrome.downloads. There is no sign-in and no extension-operated backend. Core flows: - Scan page: the content script walks the DOM for real <table> elements and for parents that have several same-tag children (heuristic “lists,” e.g. div or li grids). Optionally, if the user provided a Custom CSS selector, matches become a high-scoring “custom” dataset. Results appear as candidates in a Dataset dropdown with scores; the preview can show scan-time snapshots before a full extract. - Try next dataset: cycles the Dataset selection (Auto, then each candidate) so the user can compare which heuristic best matches the desired rows/columns before Extract or Crawl. - Custom CSS selector: querySelectorAll on the visible page; each matching element is one output row. Columns come from that element’s visible direct element children (innerText per child), or a single cell of full text if there are no such children. - Extract once: re-runs extraction on the live DOM for the chosen dataset (and custom selector if applicable) and fills the popup preview and column editors. - Crawl: repeats extract across multiple loads. Modes: (1) this page only, (2) follow “Next” links—user-located control first (saved CSS path), then heuristics (rel="next", link/button labels like Next, etc.), (3) infinite scroll—scroll to bottom between passes; if a Next/located selector is set, that control is clicked before each scroll to support “load more” style UIs. Configurable action delay, max wait, and max pages/scrolls. - Locate Next control: injects a click-to-pick overlay (nextButtonPicker.js); the clicked element’s path is stored in chrome.storage settings as nextButtonSelector (merge via background script). User can edit or clear the field manually. - Filtering: optional row filter (substring match on any column). Column names can be edited; columns can be hidden with Show checkboxes; Reset names restores headers from the last ingest. - Export: Download CSV (UTF-8 BOM), Download XLSX, or Copy TSV to the clipboard. Privacy stance in plain terms: Scraped text is processed and displayed in the extension popup and written to files the user saves; the extension does not send page content to a remote server operated by the author. Broad host permissions exist so the content script can run on pages the user chooses to scrape after a user gesture (opening the popup and using Scan/Extract/Crawl). FEATURE SUMMARY (POPUP SECTIONS AND ACTIONS) -------------------------------------------- Header: - Title: Quick Data Scraper - Tagline (sub): Scan the active tab, refine columns, export CSV or Excel. Primary panel: - Button: Scan page (primary) - Status line (dynamic; aria-live polite) Custom CSS selector (optional): - Label: Custom CSS selector (optional) - Input placeholder: e.g. article.card, [data-company] - Hint: Targets each card/root; one row per match, columns from direct children (or one column of full text). Row actions: - Button: Try next dataset (disabled until at least one dataset exists after scan) - Button: Locate Next control… Located “Next” / pagination selector: - Label: Located “Next” / pagination selector - Input placeholder: Set by “Locate…” or paste a CSS selector - Button: Clear (tiny) - Hint: Used for Follow “Next” links (runs before the automatic guess). For Infinite scroll, the control is clicked before each scroll when set. Dataset: - Label: Dataset - Select: disabled until scan finds candidates; first option is Auto (best guess); further options show labels like Table N, List (tag × rowCount), or Custom selector (rows × cols), each with a score. Timing and crawl: - Action delay (ms), Max wait (ms) - Pagination / loading: This page only | Follow “Next” links | Infinite scroll - Max pages / scrolls, Filter rows (contains) placeholder: optional Row actions: - Button: Extract once - Button: Crawl Columns panel: - Section title: Columns - Button: Reset names (tiny) - Per-column: name input, Show checkbox (generated in popup.js) Preview panel: - Section title: Preview - Button: Copy TSV (tiny) - Table: preview (first 100 rows shown in UI; export uses full filtered set) - Meta line: row/column counts and scan-preview hint when applicable (dynamic) Footer: - Button: Download CSV (primary) - Button: Download XLSX (primary) POPUP USER INTERFACE STRINGS (FOR CONSISTENCY WITH THE PRODUCT) ---------------------------------------------------------------- Static strings are in popup.html as above. Dynamic strings (popup.js / messages): Dataset dropdown: - First option text: Auto (best guess) - Other options: “${label} — score ${score}” where label comes from detection (e.g. Table 1, List (div × 19), Custom selector (N rows × M cols)). Initial status when opening popup (if no prior action): Open a web page, then click Scan. Typical post-scan status: mentions custom selector note if present, or preview line with row count and reminder to Extract once for live DOM before Crawl when showing scan preview. Other status messages are driven by errors, extract/crawl completion, clipboard copy, picker instructions, and storage updates for the saved Next selector.