TikToken + Unix Dictionary Shortener
Shorty tokenizes GitHub raw/blob URLs with TikToken, swaps every dictionary hit with a two-character marker sourced from the UNIX word list, and keeps everything on one static page. Paste a link to generate a deterministic shortcode, or drop in a shortcode/URL to restore the original download target.
Raw/Direct GitHub URL
Loading tokenizer…
Glyph Set
Unicode Box Drawing
Accepts raw URLs or github.com blob URLs (main/master supported). Encoding starts automatically.
Encoded shortcode will appear here.
Shortcode or URL
Paste a shortcode or URL to decode instantly.
Decoded URL will appear here.
How it works
The encoder splits the canonical owner/repo/@branch/path string using TikToken’s subword vocabulary. Each alphabeti…
TikToken + Unix Dictionary Shortener
Shorty tokenizes GitHub raw/blob URLs with TikToken, swaps every dictionary hit with a two-character marker sourced from the UNIX word list, and keeps everything on one static page. Paste a link to generate a deterministic shortcode, or drop in a shortcode/URL to restore the original download target.
Raw/Direct GitHub URL
Loading tokenizer…
Glyph Set
Unicode Box Drawing
Accepts raw URLs or github.com blob URLs (main/master supported). Encoding starts automatically.
Encoded shortcode will appear here.
Shortcode or URL
Paste a shortcode or URL to decode instantly.
Decoded URL will appear here.
How it works
The encoder splits the canonical owner/repo/@branch/path string using TikToken’s subword vocabulary. Each alphabetic run is checked against a cached UNIX dictionary (plus a few custom tokens); if the dictionary token is shorter, it becomes a ^page+unicode symbol that captures both the entry index and the word’s case. Decoding reverses the process by reading those markers back into words before rebuilding both the raw download URL and the standard GitHub blob link. Everything lives in index.html and words.txt, so any host can serve it without a backend.