Start here
TinyFiles in plain English
New to TinyFiles? Read this page first. No jargon, no setup yet — just what it is, what it does to your files, and the simplest way to try it. When you're ready for the full tour, jump to Getting started.
What is it?
TinyFiles (the engine is called AT-1) makes your files smaller and packs them into a single .at1 file. Two things make it different from a normal zip:
- Nothing is lost.It's “lossless,” which means when you unpack it you get the exact original bytes back — guaranteed. Every file is checked against the original before it's saved.
- You can search inside it. For tables and JSON, you can run queries directly on the small file without unpacking it first — see Query.
- It's tamper-evident. Each file carries a fingerprint, so you can prove it hasn't been changed since it was made.
Pick how you want to use it
There are three doors into TinyFiles. Most people start with the desktop app.
Drag and drop. Best if you just want to shrink some files on your computer.
1. Download the app from /download (Windows, macOS, Linux) 2. Open it, drag a file (or a whole folder) onto the window 3. You get a smaller .at1 file next to the original 4. Drag the .at1 back in any time to get your exact file back
Full guide: Desktop app · Download: /download
One install, then two commands you'll use 90% of the time:
npm i -g @tinyfiles/cli # the command-line tool — a self-contained binary, no Python required
# make a file smaller (lossless — nothing is lost) at1 compress auto report.csv report.at1 # get the original back, byte-for-byte identical at1 decompress report.at1 report.csv
Full guide: CLI reference
Call it from Python, or decode .at1 files from JavaScript, Go, Rust, C, or the browser.
Full guide: SDK & bindings
Will every file get smaller?
Honest answer: no, and that's expected. How much you save depends entirely on what the file already is:
| Kind of file | What to expect |
|---|---|
| CSV, tables, logs, JSON, sensor & financial data | Big wins — often 5–50× smaller, and searchable. |
| Databases, scientific arrays, genomics (VCF) | Big wins — purpose-built codecs. |
| Raw / uncompressed images, medical scans (DICOM) | Real wins. See Images. |
| Already-compressed files: .jpg, .png, .mp4, .zip, .gz | Little or no size win — they're already squeezed. You still get a verified, tamper-evident container. |
Not sure? TinyFiles is never worse than a plain archive — if it can't beat a standard zip on your file, it falls back to that automatically. And you can always check first with the free audit.
Do I need an account?
To unpack a file, never. Decoding any .at1 is always free and needs no login. To compress (create an archive), you connect a free account so usage can be metered — the free tier covers a one-time 100 GB-compressed trial (free account, no card). Details in Connect your account and Pricing.
Three things people ask on day one
Only files made with the queryable codecs can be searched in place: qcolumnar for tables/CSV and qjson for line-delimited JSON. Re-compress with one of those, e.g. at1 compress qcolumnar data.csv data.at1.
That's correct — JPEGs and MP4s are already compressed. TinyFiles won't re-shrink them much (nothing lossless will). It's not a replacement for JPEG. See Images for what does win.
You always do — compression is verified against the original before the .at1 is written, and at1 verify file.at1 re-checks the fingerprint any time.
One command worth knowing
Whenever you're unsure what a .at1 file is or how well it compressed:
at1 info report.at1 # what's inside, how small it got, is it queryable