Turn files into glyphs, get the exact bytes back, and take them
offline. Files stay in your browser. No uploads. No framework.
Loading the local engines…
01 Workbench
Start with a real example
Encoding options
Browser gzip is separate from the codec. Its compression level is
chosen by the browser. UTF-8 modes use their own bounded mask
search.
Bytes in
UTF-8 text
GlyphPack payload
UTF-16 code units
Choose an example to get started.
Codec input bytes—
Native payload bytes—
Same glyphs · UTF-8—
Equivalent base64—
GlyphPack is an encoding, not compression. Native size follows the
chosen UTF-8, UTF-16LE or Windows-1252 transport.
Bytes out
The decoded result will appear here.
A preview of the actual decoded bytes.
Local decode timing
Warmed median of seven batches; loading and encoding are excluded;
the selected verification is included. Small inputs are batched.
This is a local comparison, not a cross-device claim.
Use this decoder in your code
02 Small, useful recipes
Load a packed file
A .gp16 payload contains UTF-16LE code units. Read bytes before
decoding text.
import {binary} from './full.min.mjs';
const res = await fetch('asset.gp16');
const payload = new TextDecoder('utf-16le')
.decode(await res.arrayBuffer());
const bytes = binary(payload);
Restore structured data
JSON, manifests and configuration keep their exact original UTF-8
bytes.
Each file is self-contained. Native gzip + base64 loads its Wasm once
during import; binary() is synchronous afterward. The optional memory
lookup is off by default.