README

PHP library for conversion between JSON and TOON (Token-Oriented Object Notation) - an optimized format for saving tokens in Large Language Models (LLMs).

☕ Sponsors

If you find this plugin useful, consider sponsoring its development:

Sponsor on GitHub

📖 What is TOON?

TOON (Token-Oriented Object Notation) is a compact and readable serialization format, specifically designed to reduce token usage when sending structured data to LLMs. It can save 30-60% of tokens compared to JSON.

Why use TOON?

💸 Token savings: 30-60% fewer tokens than JSON
🤖 Optimized for LLMs: Better understanding and accuracy
📊 Perfect for tabular data: Uniform arrays of objects
🔄 Lossless conversion: Converts to and from JSON without data loss

Quick Comparison

JSON (123 tokens):

{
  "title": "Employees",
  "total": 3,
  "data": [
    {
      "id": 1,
      "name": "Roni",
      "email": "roni@phpiando.com"
    },
    {
      "id": 2,
      "name": "Sommerfeld",
      "email": "sommerfeld@phpiando.com"
    },
    {
      "id": 3,
      "name": "PHPiando",
      "email": "phpiando@phpiando.com"
    }
  ]
}

TOON (64 tokens - 48% savings):

title: Employees
total: 3
data[3]{id,name,email}:
  1,Roni,roni@phpiando.com
  2,Sommerfeld,sommerfeld@phpiando.com
  3,PHPiando,phpiando@phpiando.com

🚀 Installation

composer require phpiando/toonify

Requirements

PHP 8.3 or higher
ext-json
ext-mbstring

📚 Basic Usage

Simple Conversion

use Toonify\Toon;

// PHP Array to TOON
$data = [
  'title' => 'Employees',
  'total' => 3,
  'data' => [
    ['id' => 1, 'name' => 'Roni', 'email' => 'roni@phpiando.com'],
    ['id' => 2, 'name' => 'Sommerfeld', 'email' => 'sommerfeld@phpiando.com'],
    ['id' => 3, 'name' => 'PHPiando', 'email' => 'phpiando@phpiando.com'],
  ]
];

$toon = Toonify::encode($data);
echo $toon;
// title: Employees
// total: 3
// data[3]{id,name,email}:
//   1,Roni,roni@phpiando.com
//   2,Sommerfeld,sommerfeld@phpiando.com
//   3,PHPiando,phpiando@phpiando.com

// TOON to PHP Array
$decoded = Toonify::decode($toon);

From JSON String

$json = '{"name": "Roni Sommerfeld", "age": 37}';
$toon = Toonify::fromJsonString($json);

$jsonBack = Toonify::toJsonString($toon);

From Local Files

// JSON → TOON
$toon = Toonify::fromJsonDisk('/path/to/data.json');
file_put_contents('/path/to/output.toon', $toon);

// TOON → JSON
$json = Toonify::toJsonDisk('/path/to/data.toon');
file_put_contents('/path/to/output.json', $json);

From URLs

// Fetch JSON from URL and convert to TOON
$toon = Toonify::fromJsonUrl('https://api.example.com/data.json');

// Fetch TOON from URL and convert to JSON
$json = Toonify::toJsonUrl('https://example.com/data.toon');

Extract TOON from Markdown (LLM Responses)

A special feature for working with LLM responses that frequently return data in markdown blocks:

$llmResponse = <<<'MARKDOWN'
Here is the data:

```toon
users[2]{id,name}:
  1,Roni
  2,PHPiando

Hope this helps! MARKDOWN;

// Extract only TOON content $toon = Toonify::extractFromMarkdown($llmResponse);

// Or convert directly to JSON $json = Toonify::toJsonString($toon);


## ⚙️ Configuration Options

### Encoding Options

```php
$toon = Toonify::encode($data, [
  'delimiter' => ',',      // ',', "\t" or '|'
  'indent' => 2,           // Spaces per level
  'lengthMarker' => '',    // Length prefix (optional)
]);

Delimiter examples:

// Comma (default)
$toon = Toonify::encode($data, ['delimiter' => ',']);
// users[2,]{id,name}:

// Tab (more token efficient)
$toon = Toonify::encode($data, ['delimiter' => "\t"]);
// users[2	]{id,name}:

// Pipe
$toon = Toonify::encode($data, ['delimiter' => '|']);
// users[2|]{id,name}:

Decoding Options

$data = Toonify::decode($toon, [
  'strict' => true,        // Strict validation (default: true)
  'indent' => 2,           // Expected spaces per level
]);

📋 Supported Formats

Simple Objects

['name' => 'Roni Sommerfeld', 'age' => 37]

name: Roni Sommerfeld
age: 37

Primitive Arrays

[1, 2, 3, 4, 5]

[5,]: 1,2,3,4,5

Tabular Arrays (Sweet Spot!)

[
  ['id' => 1, 'name' => 'Roni'],
  ['id' => 2, 'name' => 'PHPiando']
]

[2,]{id,name}:
  1,Roni
  2,PHPiando

Mixed Arrays

[
  ['x' => 1],
  42,
  'hello'
]

[3,]:
  - x: 1
  - 42
  - hello

🎯 Use Cases

1. Send data to LLMs

$repositories = fetchGitHubRepos();
$toon = Toonify::encode($repositories, ['delimiter' => "\t"]);

// Save tokens in prompt
$prompt = "Analyze these repositories:\n\n" . $toon;
$response = $llm->complete($prompt);

2. Process LLM responses

$response = $llm->complete("List 5 products in TOON format");
$toon = Toonify::extractFromMarkdown($response);
$products = Toonify::decode($toon);

3. Optimized APIs

// Endpoint that returns TOON instead of JSON
header('Content-Type: text/plain');
$data = $database->query('SELECT * FROM users');
echo Toonify::encode($data);

4. Compact logs

$logData = [
  'timestamp' => time(),
  'events' => $events
];
file_put_contents('log.toon', Toonify::encode($logData));

🧪 Testing

composer test

📊 Benchmarks

Run examples to see token savings:

php examples/basic.php

Typical results:

Simple objects: ~20-30% savings
Tabular arrays: ~40-60% savings
Mixed arrays: ~25-35% savings

🤝 Contributing

Contributions are welcome! Please:

Fork the project
Create a feature branch (git checkout -b feature/MyFeature)
Commit your changes (git commit -am 'Add new feature')
Push to the branch (git push origin feature/MyFeature)
Open a Pull Request

📄 License

This project is under the MIT license. See the LICENSE file for more details.

🔗 Useful Links

🙏 Credits

TOON was created by Johann Schopplich.

This PHP library is an implementation of the official TOON v1.3 specification.

💬 Support

🐛 Issues: GitHub Issues
💡 Discussions: GitHub Discussions

Made with ❤️ for the PHP community

phpiando / toonify

Maintainers

Details