How do I open a .at1 file?

Three ways, none of which require installing software. Drop it into the decoder on this page and it is read locally in your browser — nothing is uploaded. Or run `npx @tinyfiles/cli decompress file.at1 out.csv`, which downloads nothing permanent and extracts any archive byte-for-byte. Or compile the public single-file C reference decoder, which works offline and forever.

Do I need to install AT-1 to read an archive someone sent me?

No. Decoding never requires an account or a licence, and it never requires a permanent install. The browser decoder on this page runs the same C block decoder used inside DuckDB and Postgres, compiled to WebAssembly, entirely on your device. Only compressing past the free tier is metered — reading is always free.

Does my file get uploaded when I open it in the browser?

No. There is no upload, no account and no backend call. The WebAssembly decoder runs in your browser and reads the file from your own disk, which is why the page works with your network disconnected. You can verify this in your browser's network tab.

Why not ship self-extracting archives that run themselves?

Because an archive that executes code cannot be evidence. AT-1 archives are used as tamper-evident records, and an auditor is right to reject a file that runs a program when opened. Self-extracting executables are also a primary malware delivery route, so mail gateways and endpoint security block them, and a signed executable would need re-signing for every single archive. A sandboxed browser decoder and a public reference decoder reach more devices with none of that risk.

What happens to my archives if TinyFiles disappears?

They stay readable. The container format is documented and there is a single-file C reference decoder in the public repository — small enough to audit, portable enough to compile anywhere. That is deliberate: a format you can only open with one vendor's software is a liability, not an archive.

Nothing to install

Someone sent you a .at1

Open it right here. It is read on your device — no upload, no account, no software.

Drop a .at1 archive here

It is read on this device. Nothing is uploaded, and there is no account.

Choose a file

Haven’t got one? (36 KB of synthetic forecourt telemetry)

Full extraction, still nothing installed

The browser decoder covers queryable archives. To extract any archive byte-for-byte, this runs it without installing anything (it needs Node, and downloads nothing permanent):

npx @tinyfiles/cli decompress your-file.at1 out.csv

In the browser

Any OS, any device, nothing installed and nothing uploaded. The decoder is the same 84 KB C core that runs inside DuckDB and Postgres, compiled to WebAssembly.

One command, no install

npx @tinyfiles/cli decompress file.at1 out.csv — extracts any archive byte-for-byte, leaves nothing permanent behind, and never asks for an account.

Forever, without us

A public single-file C reference decoder, compilable anywhere and offline. An archive you can only open with one vendor's software is a liability, not an archive.

What the browser decoder does and doesn’t do

It reads queryable, zstd-profile archives and previews integer columns. Archives written with the xz profile decode perfectly well natively but have no WebAssembly port, so the page will tell you plainly rather than fail in a confusing way. For any archive at all, the npx command above extracts it byte-for-byte without installing anything.

See the verification demo Install the CLI SDK & bindings

Common questions

How do I open a .at1 file?: Three ways, none of which require installing software. Drop it into the decoder on this page and it is read locally in your browser — nothing is uploaded. Or run `npx @tinyfiles/cli decompress file.at1 out.csv`, which downloads nothing permanent and extracts any archive byte-for-byte. Or compile the public single-file C reference decoder, which works offline and forever.
Do I need to install AT-1 to read an archive someone sent me?: No. Decoding never requires an account or a licence, and it never requires a permanent install. The browser decoder on this page runs the same C block decoder used inside DuckDB and Postgres, compiled to WebAssembly, entirely on your device. Only compressing past the free tier is metered — reading is always free.
Does my file get uploaded when I open it in the browser?: No. There is no upload, no account and no backend call. The WebAssembly decoder runs in your browser and reads the file from your own disk, which is why the page works with your network disconnected. You can verify this in your browser's network tab.
Why not ship self-extracting archives that run themselves?: Because an archive that executes code cannot be evidence. AT-1 archives are used as tamper-evident records, and an auditor is right to reject a file that runs a program when opened. Self-extracting executables are also a primary malware delivery route, so mail gateways and endpoint security block them, and a signed executable would need re-signing for every single archive. A sandboxed browser decoder and a public reference decoder reach more devices with none of that risk.
What happens to my archives if TinyFiles disappears?: They stay readable. The container format is documented and there is a single-file C reference decoder in the public repository — small enough to audit, portable enough to compile anywhere. That is deliberate: a format you can only open with one vendor's software is a liability, not an archive.