Parses your TOON via a full round-trip through the data model, then re-serialises it cleanly. Catches structural errors, malformed tabular arrays, and invalid scalar values.
Choose comma or pipe as the tabular delimiter and set 2 or 4-space indentation for nested objects. Useful for re-encoding TOON to match a specific LLM prompt style or downstream parser.
Uniform arrays of objects are automatically collapsed into the compact tabular header-plus-rows format, maximising token savings when sending data to an LLM.
All formatting runs in your browser — no data is sent to any server. Safe for TOON payloads containing business data, PII, or proprietary schemas.
TOON (Token-Oriented Object Notation) is a compact, lossless encoding of the JSON data model designed to reduce token count when sending structured data to large language models. It uses YAML-style indentation for objects and a CSV-style tabular layout for uniform arrays, achieving up to 60% fewer tokens than equivalent JSON.
Consistent formatting helps LLMs parse structure reliably. The formatter ensures tabular arrays are properly collapsed, indentation is uniform, and all scalar values are correctly quoted — reducing the chance of the model misreading the structure.
When all objects in an array share the same primitive-valued fields, TOON writes a header key[N]{f1,f2,f3}: followed by one comma-separated data row per object. This is the biggest source of token savings compared to JSON.
Yes. The formatter works by parsing TOON into the full JSON data model and then re-encoding it. All values, types, and structure are preserved. The only changes are whitespace normalisation and delimiter/indent adjustments you explicitly choose.