Hi all,
I’m releasing Tribit, a 36 bit symbolic compression system that encodes over 444,253 unique words into deterministic 6×6 bitmaps. This enables full token symbolic replacement, extreme compression, and novel AI compute interaction models.
What Tribit Enables:
Direct glyph based token substitution
Command line compression: replace strings of shell/code logic with single symbols
Massive tokenizer simplification and compression for AI models
Visual representation of model logic or indexed data
High density looped AI memory systems (persistent self writing bots)
What’s Included in the Dev Kit:
Full TTF font pack (444,253 glyphs)
Encoder/decoder scripts
JSONL source vocabulary
Legal metadata, license terms, and batch installers
Symbol previews and font demos
Access It Here:
Use Cases:
Custom tokenizer training
Memory loop reasoning in small models
Hardware-level symbolic interfacing
CLI command abstraction
Visual language construction
Dataset compression experiments
I’m opening this for feedback and experimentation. The Dev Tier is free. If anyone’s working on symbolic logic systems, persistent memory models, or post token architectures, I’d love to connect.
Andy, Founder of Triskel Data PTY LTD