toon: what it is, what problem it solves & why it's gaining traction

What it solves

TOON (Token-Oriented Object Notation) reduces the token cost of providing structured data to Large Language Models (LLMs). Standard JSON is often too verbose, leading to higher costs and potentially lower efficiency in large context windows. TOON provides a compact, lossless representation of the JSON data model that is easier for models to follow.

How it works

TOON acts as a translation layer for JSON. It combines YAML-like indentation for nested objects with a CSV-style tabular layout for uniform arrays of objects. By declaring array lengths [N] and field headers {fields} once, it eliminates the need to repeat keys for every item in a list, significantly reducing token usage while maintaining explicit structure.

Who it’s for

Developers building LLM-powered applications who need to pass large amounts of structured data into a model's prompt without incurring excessive token costs or losing structural reliability.

Highlights

• Token Efficiency: Uses approximately 40% fewer tokens than JSON in mixed-structure benchmarks.
• JSON Compatibility: A lossless, drop-in representation of the JSON data model.
• LMM-Friendly Guardrails: Includes explicit array lengths and field headers to improve parsing reliability.
• Multi-Language Support: Spec-driven implementations available in TypeScript, Python, Go, Rust, and .NET.
• Tabular Arrays: Collapses uniform arrays of objects into compact tables.