Show HN: Slim – 50% fewer tokens than JSON for LLM applications

I built SLIM because JSON wastes tokens on repeated keys. Every object in an array repeats "id", "name", "email"...

SLIM defines the schema once, then just data rows:

JSON: {"id":1,"name":"Alice"},{"id":2,"name":"Bob"} SLIM: |id#,name$|1,Alice|2,Bob

Results: 50%+ token savings, still human-readable, LLMs understand it without parsing.

MIT licensed, zero dependencies, TypeScript.

Would love feedback from the HN community.