Changelog · every release · honest

Everything we've shipped.

csvtodashboard runs 100% in your browser — no accounts, no uploads — so a release here means the site itself got better. The dates below are extracted from the project's git history; the notes are written by hand, not dumped from commit messages.

V10.5 ·

Quality rules with evidence

  • CSV Profiler now separates observed findings from explicit pass/fail contract checks
  • Reusable data-quality contracts cover required, type, unique, range, pattern, allowed-value, and composite business-key rules without storing source data
  • Row-level failure evidence identifies blank and duplicate keys, invalid values, and unavailable contract columns while aggregate reports exclude raw values
  • Column profiles now include visible completeness denominators, distinct rates, mixed value shapes, exact row duplicates, date ranges, quantiles, and IQR outliers
  • A new CSV data-quality audit guide documents profiling, contracts, thresholds, evidence, and recurring handoff
  • Standalone CSP-locked HTML, safe Markdown, complete profile CSV, and explicitly disclosed issue CSV exports keep each audit portable
V10.4 ·

Variance that explains itself

  • CSV Compare now builds a complete baseline-versus-current report with exact metric deltas and denominator-safe percentage changes
  • Category drivers rank the groups behind a movement and show each contribution to net change, including new and disappeared categories
  • Schema drift and missing-value drift expose structural and completeness changes before they become reporting surprises
  • Reusable setups save only labels and selected column names — never filenames, rows, totals, deltas or evidence
  • A new CSV variance analysis guide documents the calculation and evidence contract
  • Standalone HTML, Markdown and full driver CSV exports keep the calculations portable; optional data-in-link sharing requires an explicit disclosure
V9 ·

Sharper pages, honest by default

  • Save a cleaning pipeline as a named profile and re-run it on any new file in one click — build once, reuse forever
  • Every “CSV to X” page now leads with a direct answer, the format's real gotchas and a worked example — not boilerplate
  • A live privacy indicator on the homepage: flip on airplane mode and watch it switch to “Offline” while every tool keeps working
  • Clearer, clearly-labelled ad placement with a text-contrast fix for readability
V8 ·

The hardening release

  • Faster first load — the heavy file-conversion code now loads on demand, trimming about 100KB off the homepage
  • Site-wide accessibility pass — proper page landmarks, labelled form controls and stronger text contrast
  • Your data's types survive conversion — ZIP codes, leading zeros and long IDs stay exact in Excel, Arrow and Parquet exports
  • Smarter delimiter detection for European CSVs that use semicolons inside quoted fields
  • 12 more ready-made dashboard templates — Asana, Trello, Monday, Pipedrive, Airtable, Gumroad, Substack and more
  • Richer result tables in the join tool, plus better internal navigation between guides and lessons
V7 ·

The depth release

  • Analysis results are real tables now — sortable, with correlation heatmaps, outlier severity badges and in-cell data bars
  • Word-level highlighting in CSV diff — see exactly which words changed
  • Editor: freeze columns, drag to reorder, match-by-match find & replace, per-column stats, copy a selection as Markdown or JSON
  • SQL workbench — add a second file and JOIN across both, syntax highlighting, query history and saved snippets
  • Dashboards: per-chart colors, KPI trend badges, an automatic category×category heatmap, PDF export
  • Excel sheet name / filter row / fitted widths, Markdown alignment, Parquet compression — and the pipeline gains clean-headers, trim, fill-blanks and select-columns steps
  • Fixed: columns with “duration” in the name no longer render as percentages
V6 ·

The workbench release

  • Opt-in multi-file workspace — load several files once, then open any of them from any tool
  • Ctrl+K command palette across 190+ tools — switching tools carries your file with you
  • Shareable pipeline recipes and a recipes gallery
  • Mobile nav menu and a compact cookie notice
  • Accessibility pass — focus traps, reduced-motion support, skip link
  • Per-page Open Graph cards
  • Engineering notes and this changelog
  • Performance budget gate in CI
V5 ·

Compose, teach, win the click

  • Multi-key JOIN, VLOOKUP-style lookup and append across files
  • Interactive SQL-on-CSV course — 10 lessons, self-checking, fully local
  • Chart palettes, fixed-theme and TABLE embeds, plus an embed gallery
  • Cleaning suggestions on the dashboard
  • Natural-language → SQL, template-based and computed locally
  • PWA “Open with” file handlers
  • Alternatives hub — five honest comparisons
  • Three new Excel-pain guides; analytics verticals grow to 31
V4 ·

Own the head terms

V3 ·

Depth and scale

  • DuckDB Power mode for real SQL on the query tool
  • 250MB off-thread parsing engine with live progress
  • Fuzzy dedupe, outliers, percentiles, correlation, crosstab, resample and friends
  • CSV glossary, the Excel-pain guide wave and the vertical-dashboards wave — the site passes ~198 pages
V2 ·

The toolkit grows

  • 27 new tools across four batches
  • Converters in both directions — YAML, XML, TOML, INI, NDJSON, fixed-width and more
  • Cleaners, extractors, generators and formatters
  • QR codes (single and bulk from CSV) and PDF export
V1 ·

Launch

  • The CSV → dashboard builder — drop a file, get charts, KPIs and a table
  • Core converters: JSON, SQL, Excel, Markdown, HTML, XML
  • Validator, dedupe, sort & filter, split, SQL query and the pipeline builder
  • 100% client-side — no signup, no upload