Read the guideOpen Jason

The spreadsheet for files that do not fit.

Open Jason free Inspect benchmarks
  • No install — it runs in a browser tab
  • No upload — no server ever receives the file
  • No memory ceiling — paged through local storage, never held in RAM
Largest measured file20.97 GB888 million cells
Measured first import3.9 min20.97 GB CSV
Measured cold open111 msmedian after import
Local project size0.795 GiB24.6× smaller than source
Source formats

One local workspace for data and documents

Start with delimited text for free. Pro adds the Local Data Suite, including typed data files, document-shaped rows, and multi-file collection imports.

Free

Delimited text

CSV, TSV, and TXT, including gzip or single-file ZIP containers.

Pro / Local Data Suite

Structured data

JSON Lines, XLSX, SQLite, Parquet, Arrow IPC / Feather v2, dBase (.dbf), and Avro.

Pro / Local Data Suite

Documents and collections

Word (.docx) outlines or tables, PowerPoint (.pptx) slides, and EPUB content blocks. Merge up to 500 CSV tables or Word outlines; Word rows retain their source filename.

Open the file your spreadsheet refuses

Before

“File too large to open.”

After

A local project you can reopen in milliseconds.

Importledger-2026.csv
Read source20.97 GB
Parse rows148,003,112
Infer types6 columns
Write local project0.795 GiB
ELAPSED 2:24NETWORK 0 BSOURCE STAYS ON DISK

A 20.97 GB CSV became a working project in 3.9 minutes on the measured run.

CapabilitiesStreaming CSV importLocal project formatType inferenceImport limits

Move through a billion cells without waiting

A one-billion-cell project scrolled with zero dropped frames and reopened in 93 ms median.

CapabilitiesCanvas grid rendererSearch and sortFormulas and calculated columnsMeasured runs
reconciliation-2026.jason1,000,000,000 cellsD4:D999999999=SUM(D:D)
ABCDEF
1txn_idposted_ataccountamountfeebalance
24180029982026-07-02ACC-103719.730.37814.47
34180030052026-07-03ACC-107427.060.511117.06
44180030122026-07-04ACC-111134.390.651419.65
54180030192026-07-05ACC-114841.720.791722.24
64180030262026-07-06ACC-118549.050.932024.83
74180030332026-07-07ACC-122256.381.072327.42
84180030402026-07-08ACC-125963.711.212630.01
94180030472026-07-09ACC-129671.041.352932.60
104180030542026-07-10ACC-133378.371.493235.19
114180030612026-07-11ACC-137085.701.633537.78
124180030682026-07-12ACC-140793.031.773840.37
134180030752026-07-13ACC-1444100.361.914142.96
144180030822026-07-14ACC-148117.690.34730.26
154180030892026-07-15ACC-151825.020.481032.85
SUM 41,882,004.19ROWS 166,666,666LOCAL PROJECT — NOT UPLOADED
Rendered per frame

~2,400

visible cells — the other billion stay in the local project until you scroll to them.

Keep every row on your own machine

Bytes uploaded

0

Import, edit, calculate and export all happen inside the browser tab.

Data boundary0 rows leave the device
Your machine
ImportSearchSortEditCalculateExport
Jason servers
Account emailSubscription stateSigned licence receipt
DATA PLANE LOCALCOMMERCIAL PLANE MINIMAL

Spreadsheet content never reaches a Jason server — there is nothing to upload and nothing to breach.

CapabilitiesData boundaryLicense modelOffline behaviourExport to CSV

Why not just use one of the others

Three steps, and only the last one is a claim about anyone else — so that is the one with a number attached.

01

Every local option is the same engine

The browser tools that open big files without uploading them are DuckDB-WASM underneath. That is why they read the same formats: it is the same reader. It is also why they share the same limit — the data lives in memory, and we measured it running out at 1.76 GB.

See the measurement
02

So the difference is not format. It is scale.

Adding one more file extension changes nothing about whether you can work in the file once it opens. Jason pages the data through local storage instead of holding it in RAM, which is a different answer to a different question.

03

No install, no upload, no ceiling

Desktop tools drop the first. Cloud tools drop the second — your file goes to their server. Local browser tools drop the third. Jason is the combination: nothing to install, nothing uploaded, and measured holding 11.78 GB and a billion cells open in a tab.

See the measurement

One file, one working loop

01

Open

Import CSV for free. Pro adds structured-data and document sources through the Local Data Suite. Nothing is uploaded to Jason.

02

Work

Search, sort, edit, calculate, and navigate the full dataset through a responsive canvas grid.

03

Return

Reopen the local project without paying the original import cost again.

04

Export

Write the result to CSV, or export the current view to Parquet with Pro.

Recently shipped

Benchmark

Billion-cell scroll run

11.78 GB source imported in 174.6 s, then scrolled with zero dropped frames.

Engine

Exact accumulation for money columns

Sum and average over currency columns no longer drift on long columns.

Performance

Cold open under 120 ms

Reopening an imported project no longer repeats the import cost.

Security

Signed license receipts

Pro capability is granted only against a subject-bound ES256 receipt.

Stop shrinking the question to fit the tool.

Open an oversized CSV for free. Pro removes Jason's limits and adds structured-data, document, and multi-file Local Data Suite tools.

Open Jason free Inspect the benchmarks