Efficiency for AI
X-Ray: .TOON
Token-Oriented Object Notation
context:
location: Boulder
hikes[3]{id,name,companion}:
1,Blue Lake Trail,ana
2,Ridge Overlook,luis
3,Wildflower Loop,sam
YAML Structure
Indentation hierarchy.
Saves token braces {}.
Field Header
Defines [N] elements and {fields} only once.
Tabular Data
CSV Style. Optimized for
long lists of objects.
Purpose
Designed to be the ideal input for LLMs. Replaces JSON to reduce costs and improve model accuracy.
Key Fact
Achieves up to ~40% token savings compared to JSON, while maintaining 74% accuracy.
Compact
Removes unnecessary quotes, braces and brackets.
Guardrails
[N] count helps the model not skip data.
Lossless
Perfect round-trip conversion with JSON.
LLM-First
Designed for AI to "understand" the structure better.
xeland314 .TOON saves • LLMs appreciate.