Release notes

Changelog

Track what changed in TableSnap, what is current right now, and where recent releases made table capture and export workflows sharper.

Current version

v0.5.0

Released November 6, 2026. Use TableSnap in 25 interface languages, with right-to-left support and localized Chrome Web Store listings.

Released November 6, 2026Current version
Added

Seven additional interface locales

Added complete Arabic, Hebrew, Hindi, Thai, Danish, Swedish and Latin American Spanish translations across the popup, preview, settings, page controls and notifications.

Improved

Improved

Regional Spanish browser languages now select the Latin American Spanish interface where appropriate; legacy Hebrew language codes are recognized.

Released

v0.5.0

Use TableSnap in 25 interface languages, with right-to-left support and localized Chrome Web Store listings.

Added

Seven additional interface locales

  • Added complete Arabic, Hebrew, Hindi, Thai, Danish, Swedish and Latin American Spanish translations across the popup, preview, settings, page controls and notifications.
  • Added localized Chrome Web Store titles, summaries, detailed descriptions and screenshot captions, bringing the store package to 26 locales.
  • Added right-to-left layouts for Arabic and Hebrew while preserving table column order and the direction of editable cell content.

Improved

  • Regional Spanish browser languages now select the Latin American Spanish interface where appropriate; legacy Hebrew language codes are recognized.
  • Plural messages correctly handle counts formatted with Arabic, Persian, Devanagari and Thai digits.
  • Existing localized store descriptions list all 25 interface languages. Image recognition continues to use the 18 bundled OCR language models, including Spanish for both Spanish interface variants.
Released

v0.4.0

Use TableSnap in 18 interface and OCR languages, with seven new fully localized Chrome Web Store listings.

Added

Seven new languages

  • Added complete Dutch, Indonesian, Vietnamese, Czech, Korean, Simplified Chinese and Traditional Chinese interfaces across the popup, preview, settings, on-page controls, notifications and accessibility announcements.
  • Added localized Chrome Web Store titles, summaries, detailed descriptions and screenshot captions for every new language.
  • Added local OCR models for all seven languages, including separate Simplified and Traditional Chinese recognition.

Improved

  • Locale detection now distinguishes Simplified and Traditional Chinese from browser region and script codes, including zh-Hans, zh-Hant, zh-CN, zh-TW, zh-HK and underscore variants.
  • Language-specific plural forms, placeholders and drag-and-drop announcements are validated automatically for every new locale.
  • Store descriptions and existing language lists now describe all 18 supported interface and OCR languages.
Released

v0.3.0

Turn tables inside images into editable, export-ready data with private, on-device OCR in every TableSnap language.

Added

Local OCR for image tables

  • Choose **OCR area**, drag around a visible table in an image, canvas, scan or screenshot, and open the recognized cells in the regular editable preview.
  • OCR supports English, Ukrainian, Polish, German, French, Spanish, Italian, Portuguese, Turkish, Japanese and Russian, selected from the page and interface languages with English support for mixed codes and numbers.
  • Recognition runs entirely inside the extension using bundled language models. The selected image is never uploaded or sent to a remote service.

Improved

  • Adaptive OCR retries use rotated, grayscale and binary representations when the first pass is weak or contains obvious artifacts.
  • Multiple OCR passes are combined cell by cell so a cleaner result cannot discard columns or overwrite cells that were already correct.
  • Table reconstruction now handles aligned and right-aligned columns, low-contrast and dark tables, captions above tables and common numeric lookalikes.
  • Japanese segmentation spaces and invisible OCR control characters are removed automatically.
  • OCR selections include a small white border to improve recognition near cropped table edges.
Released

v0.2.0

Copy and download any web table into Excel, Google Sheets, CSV or Markdown exactly as the page shows it — with real .xlsx downloads, reliable one-click copy and an interface in 11 languages.

Added

Excel .xlsx downloads

  • Download any table as a real .xlsx workbook with a bold, frozen and filterable header row, numbers and dates stored as values, and IDs with leading zeros kept as text. No row limit.

Open in Google Sheets

  • Copy a table and open a new Google Sheet in one step, from the popup or the preview, then paste with Ctrl/Cmd+V.

Feedback for every copy

  • Every copy from the page toolbar, the right-click menu, the keyboard shortcut or a manual pick now shows a short confirmation with the row and column count, or the reason it failed.

Interface in 11 languages

  • English, Ukrainian, Polish, German, French, Spanish, Italian, Portuguese, Turkish, Japanese and Russian are now complete, with correct plural forms. The store name and description are localized too.

Improved

Copies that paste straight into cells

  • CSV and TSV copies now also carry a table, so Excel, Google Sheets, Numbers and Notion paste into cells, while text editors still get plain CSV/TSV.

Cells exactly as the page shows them

  • Text hidden by the page, screen-reader duplicates and footnote markers like [1] are left out; icons keep their meaning ("Available"), checkboxes and inputs keep their values, and lists keep their line breaks.

Table structure on real sites

  • Multi-row headers are combined, repeated header rows and footnote rows are dropped, data grids are read in the right order, selection checkbox columns are removed, and partly loaded virtual grids say how many rows are loaded.

Smarter detection

  • Layout tables and card grids are no longer listed as tables, data tables nested inside layouts are found, and a table you pick by hand is always kept.

A preview you can trust

  • The preview no longer changes numbers, dates or yes/no values unless you choose a column type, cleaning options apply right away, edits survive when the source tab closes, and "Filtered" copies every matching row.

Works right after install and updates

  • Tabs that were open before installing or updating TableSnap work without a reload, and pages TableSnap cannot read (Chrome pages, the Web Store, PDFs, local files) explain why.

Faster and lighter

  • Scanning reads only each table's headers and first rows, pages without tables are not rescanned, the on-page script is about a third of its former size, and the package is less than half as large.

Accessibility

  • Visible focus everywhere, named grid cells, announced status messages and keyboard support across the popup, preview and settings.

Fixed

  • Copy as CSV, TSV and JSON failed in Chrome; it works again from every entry point.
  • Large numbers such as 8,336,817 or decimals like 1.080 could be rewritten on copy or download.
  • Spreadsheet formulas in a page could run in Excel after pasting; formula-like cells are now protected, while ordinary values stay unchanged.
  • Web pages could trigger TableSnap's toolbar actions; only your own clicks and keys count now.
  • "Auto-detect tables after page load" did not fully stop page scanning; turning it off now does.
  • "Capture + next" could press form or delete buttons; it now only follows real pagination.
  • "Load more" lists no longer duplicate rows in a collection.
  • Keyboard shortcuts no longer interfere with typing in text fields.

Privacy

  • The privacy policy and settings now say exactly what TableSnap reads and when. Everything still runs locally; nothing is sent anywhere.
  • TableSnap no longer requests the activeTab permission and no longer exposes its files to websites.

Notes

  • JSON keys from number-like headers now start with "_" (for example "_2024"), and duplicate headers get the next free suffix.
Released

v0.1.1

Focused this release on preview reliability, calmer editing UX, and a public-facing release site.

Added

Spreadsheet-like editing

  • Added multi-cell paste into preview grids with support for tabular clipboard blocks, automatic draft growth, and undo.

Local selection actions

  • Added a sheet-level selection action bar with quick copy, delete, and clear actions next to the table.

Public release site

  • Added the public marketing site, changelog page, support pages, legal pages, and blog foundation for SEO and GEO discoverability.

Improved

Preview workspace polish

  • Refined the full-window preview with a sheet-style canvas, quieter hover controls, collapsible secondary panels, inline row and column add actions, and clearer visual hierarchy.

Drag-and-drop editing

  • Reworked row, column, and collection reordering on top of dnd-kit with safer collision handling, better insertion markers, keyboard support, and more predictable drag states.

Editing feedback

  • Added toast feedback, undo actions, and spreadsheet-like keyboard navigation for row and column editing flows.

Fixed

Collection reliability

  • Fixed unstable collection item IDs and content-blind deduplication so merged tables no longer collide or silently drop distinct pages with the same shape.

Capture and shortcut behavior

  • Fixed site-specific format handling for shortcuts and context menu actions, tightened Capture + next pagination targeting, and added safer background messaging behavior on restricted or unready tabs.
Released

v0.1.0

Added automatic detection for semantic HTML tables, ARIA grids, and supported custom layout tables.

Added popup workflows for copy, preview, quick export actions, and collection building.

Added full-window preview with editing, drag-and-drop row and column ordering, filtering, and export presets.

Added CSV, TSV, Markdown, HTML, and JSON output with Excel, Sheets, Notion, and Airtable-friendly options.

Added local-only settings, site-specific recipes, transform tools, and merge workflows for multi-table capture.