Open a million-row CSV, Excel, JSON or XML file in seconds. Clean it, reshape it, chart it, and query it with real SQL — all on your own computer. Nothing is ever uploaded.
Windows 10 & 11 · One-time purchase, all features, no subscription, no account · macOS & Linux planned
Spreadsheets mangle IDs, ZIP codes and dates, and slow to a crawl past a few hundred thousand rows. Online CSV tools make you upload sensitive data to someone else's server. Database GUIs are overkill for a one-off cleanup job. MountainCSV sits in the gap: the speed and SQL power of a real analytical database, the approachability of a spreadsheet, and the privacy of a local app.
A real embedded columnar engine under the hood — not a script bolted onto a CSV parser. Smooth scrolling through millions of rows; only the visible page is ever held in memory.
Write actual SQL against every open file — joins, window functions, CTEs, aggregations — with zero setup, and the result opens as a new tab.
No telemetry, no analytics, no network calls, no account. Ideal for PII, financials, and anything under NDA or GDPR.
No hunting through nested ribbons. One click switches the whole screen between the three things you actually do with data — and in Dashboard and SQL your data stays right beside you, so you never lose sight of what you're working on.
Your grid, full screen, with spreadsheet column letters and row numbers. Search, filter, find & replace, conditional formatting, type conversion, validation and in-cell formulas — everything for getting a file into shape, and nothing else in the way.
A real SQL editor with schema-aware autocomplete and a browsable table tree, plus every reshaping tool — Join, Union, Compare, Aggregate, Pivot, Correlate, Unpivot and Sample — on one toolbar. Results open as a new tab in the data panel beside you.
Not a SQL person? Use the dialogs, then hit Edit as SQL to see the exact query they'd run — and change it. It's the fastest way to learn SQL on your own data.
Build from KPI, KPI-strip, chart, table, frequency, slicer and text tiles — or press Auto and get a sensible dashboard of your file in one click. Click a slicer value or a chart bar and every tile filters together, Power-BI style. Six chart types (bar, line, pie, scatter, histogram, map) with PNG/SVG/CSV export; maximise any chart to fill the window; publish the board to a self-contained HTML page or PDF. And your dashboard is saved per file — reopen the same file next week and it's exactly as you left it.
A full data-cleaning and analysis workbench, not just a CSV viewer.
CSV, TSV, and other delimited text; Excel (.xlsx, .xls, .xlsm,
.xlsb) with a sheet picker; JSON & NDJSON; XML. Automatic encoding detection
(UTF-8/16, ISO-8859, Windows-125x, GBK, Shift-JIS) with one-click override, and an Import Preview
that shows the detected delimiter, encoding and column types before you commit.
customer.address.city) to a real column in one click
Every change is undoable (Ctrl/⌘+Z) — including the destructive ones. Edit cells inline,
use Excel-style in-cell formulas (=A1+B1, SUM(A1:A10)), add computed
columns with live-previewed SQL expressions, and pick columns to reference from clickable
chips — never type or misremember a column name.
🏆 Cluster & merge automatically groups values that mean the same thing ("NYC" / "New York" / "new york") so you can collapse them in one click — on an engine that doesn't choke past a million rows. Per-column validation rules (required, integer, decimal, min/max, regex, email, URL, unique, date, one-of-list) with one-click auto-fixes.
Pivot and unpivot, aggregate with GROUP BY, join two files on a key, or 🏆 fuzzy-join on approximate keys (Jaro-Winkler) so "Jonathan" matches "Jonathon" — an enterprise feature most tools charge hundreds for. Union multiple files, or sample a random subset into a new tab.
Correlation matrix, duplicate detection, outlier detection (IQR and Z-score). Interactive charts — bar, line, pie, scatter, histogram and map — with cross-filtering: click a bar, slice or histogram bin and the data beside it filters instantly, with a chip showing exactly what's applied and a single click to undo it.
🏆 Find in Files searches a whole folder of data files without opening any of them — "which of these forty exports mentions order 1001?" — and lands the matches (file, row, column, value) as an ordinary tab you can sort, filter and export.
The standout feature: a full SQL workspace, not a cramped popup. Every open file is exposed as a named table — write joins, window functions, CTEs and aggregations, and the result opens as a new tab beside the editor. Schema-aware autocomplete, syntax highlighting, a browsable table tree and a persisted query history.
Federated SQL — attach a live database connection and query its tables and your files together in one statement: join this morning's CSV export straight onto the database table it came from, no export step in between. Filters and column selections are pushed down to the database, so only the rows you ask for cross the wire.
SELECT region, SUM(revenue) AS total
FROM sales
JOIN regions ON sales.region_id = regions.id
GROUP BY region
ORDER BY total DESCA file lands on your desk and the question is always the same: does this match what's actually in the system? The usual answer is a staging table, a load script, and permissions you don't have.
Attach the database — SQLite, PostgreSQL, Oracle, or anything with an ODBC driver — and join the file to the live table in one statement. No import, no temp table, no CREATE privileges. Filters and column selections are pushed down to the server, so only the rows you asked for ever cross the wire, and the file itself never leaves your machine.
Or skip the SQL entirely: Compare takes two tables and reports what's only on the left, only on the right, changed, and identical.
No telemetry, no analytics, no network calls, no account, no cloud sync. MountainCSV runs entirely on your computer — a permanent "Local-only" indicator sits in the status bar so you never have to wonder. That makes it a genuine option for PII, financial data, medical records, or anything under NDA or GDPR that you simply can't upload to a browser tool. That isn't an edge case — it describes most of the data people are paid to work with.
| Spreadsheets | Online CSV tools | MountainCSV | |
|---|---|---|---|
| Million-row files | 🐌 slow / capped | varies | ⚡ Fast |
| Preserves IDs, ZIPs, dates | ❌ mangles | varies | ✅ Yes |
| Real SQL on your data | ❌ | rarely | ✅ Yes |
| Join / pivot / validate | partial | partial | ✅ Yes |
| Search a whole folder of files | ❌ | ❌ | ✅ Find in Files |
| Version safety net on save | ❌ | ❌ | ✅ Last 10 versions |
| Data stays on your machine | ✅ | ❌ uploads | ✅ Yes |
| Price | subscription | subscription | One-time £19.99 |
No subscription, no per-seat licensing, no feature gating. Buy once, use it for as long as you like.
MountainCSV is on the Microsoft Store — £19.99 once, no subscription, with a free 7-day trial of the full app. Installing from the Store means no security warnings and updates arrive automatically.
Windows 10 and 11. Your trial starts when you first open the app.
No. MountainCSV makes zero network calls with your data. Everything — opening, editing, SQL queries, charts, exports — runs locally against a database engine embedded in the app itself.
Nothing is held back during the trial — you get the whole app. When it ends, buy it from the Microsoft Store and carry on; your files, settings and recent-file list are exactly as you left them.
No — £19.99 is a one-time purchase that unlocks every feature, permanently. There's no recurring charge and no account to manage.
MountainCSV is built on an embedded columnar analytical engine, not a spreadsheet grid — it's designed for files with millions of rows. Only the visible page of data is ever held in memory.
The User Guide walks through all three workspaces and every panel and dialog in the app, step by step, with screenshots.