TOON for .NET

Encodes .NET values to TOON (Token-Oriented Object Notation) and decodes TOON back. TOON is a compact, indentation-based encoding of the JSON data model for LLM input.

Installation

dotnet add package Toon.Format

Usage

using Toon.Format;

var data = new
{
    users = new[]
    {
        new { id = 1, name = "Ada", role = "admin" },
        new { id = 2, name = "Bob", role = "user" }
    }
};

string toon = ToonEncoder.Encode(data);
// users[2]{id,name,role}:
//   1,Ada,admin
//   2,Bob,user

var node = ToonDecoder.Decode(toon);
// {"users":[{"id":1,"name":"Ada","role":"admin"},{"id":2,"name":"Bob","role":"user"}]}

ToonDecoder.Decode returns a JsonNode; ToonDecoder.Decode<T> deserializes into T through System.Text.Json. Pass a ToonEncodeOptions or ToonDecodeOptions as the second argument:

Option	Default	Description
IndentSize	2	Spaces per indentation level (encode and decode)
Delimiter	ToonDelimiter.Comma	Array delimiter: Comma, Tab, or Pipe (encode)
Strict	true	Throws on count mismatches, duplicate keys, tab or misaligned indentation, blank lines inside an array or keyed tabular object, and depth jumps; false recovers from these five only (decode)

Specification

Targets TOON spec v4.4, and the test suite runs the spec's conformance fixtures.

• Integers in long range decode to long, every other number to double – integers beyond long range lose precision and a token beyond double range (e.g. 1e999) decodes as a string (§4)
• Every .NET numeric type encodes as a number – integers in long range stay exact, ulong values above long.MaxValue and decimal values round to the nearest double, float keeps its shortest digits (0.1f encodes as 0.1), NaN and ±Infinity become null, JsonNode and JsonElement values encode as the JSON they hold, dictionaries become objects with string keys, other enumerables become arrays, and any other value encodes as System.Text.Json serializes it, honoring its attributes and converters so Decode<T> reads it back – a DateTime becomes an ISO 8601 string, an enum its number (§3)

Releases follow SemVer: a new spec MINOR version ships as a MINOR release, even when it changes how hand-written input decodes, and a MAJOR release means an API break or a new spec MAJOR version.

Migrating from spec v3.0

Spec v4 dropped key folding and path expansion, along with the KeyFolding, FlattenDepth, and ExpandPaths options, and reads a line whose first non-space character is # as a comment. Decode documents written with folded keys or #-leading lines using a v3.0 build (with ExpandPaths = ToonPathExpansion.Safe for folded keys), then encode the result again with this one.

Resources

• Specification: SPEC.md – Normative rules and conformance checklists
• Format Overview: toonformat.dev – Every form with examples
• Other Implementations: toonformat.dev – TOON in other languages

Contributing

See CONTRIBUTING.md for the development setup and pull request guidelines.

License

MIT License © 2025-PRESENT Daniel Destouche and Johann Schopplich