High-performance data transformation stack for JSON, YAML, TOML, INI, CSV, XML, MessagePack, and TOON.

🏆 415x faster YAML | 9x faster TOML | 2-3x faster INI | 30-60% LLM token savings

📦 npm | 🐙 GitHub | 🐦 X

# Install individual packages
bun add @sylphx/molt-json
bun add @sylphx/molt-yaml
bun add @sylphx/molt-toml
bun add @sylphx/molt-ini
bun add @sylphx/molt-csv
bun add @sylphx/molt-xml
bun add @sylphx/molt-msgpack
bun add @sylphx/molt-toon

# Or install all at once
bun add @sylphx/molt

import { molt } from '@sylphx/molt-json'

const data = molt(`{
  user: 'alice',        // ✅ Unquoted keys
  joined: new Date(),   // ✅ Date preserved
  id: 123n,            // ✅ BigInt preserved
}`)

import { molt } from '@sylphx/molt-yaml'

const config = molt(`
app: MyApp
database:
  host: localhost
  port: 5432
`)

import { molt } from '@sylphx/molt-toml'

const config = molt(`
[database]
host = "localhost"
port = 5432
`)

import { parseCSV } from '@sylphx/molt-csv'

const data = parseCSV('name,age,active\nAlice,30,true', {
  parseTypes: true  // Auto-detect types
})

import { molt } from '@sylphx/molt-xml'

// Handles dirty XML automatically
const data = molt('<user name=alice age=30/>', {
  cleanDirty: true
})

import { molt } from '@sylphx/molt-ini'

const config = molt(`
[database]
host = localhost
port = 5432
`)

console.log(config.database.port) // 5432

import { encode, decode } from '@sylphx/molt-msgpack'

const data = { user: 'alice', id: 123 }
const binary = encode(data) // 20-50% smaller than JSON
const restored = decode(binary)

import { stringify } from '@sylphx/molt-toon'

const data = {
  users: [
    { id: 1, name: 'Alice', age: 30 },
    { id: 2, name: 'Bob', age: 25 }
  ]
}

// 30-60% fewer tokens for LLM prompts!
const toon = stringify(data)
// users:
//   id | name  | age
//   1  | Alice | 30
//   2  | Bob   | 25

See BENCHMARKS.md for complete results.

Key Advantages:

• 🥇 Fastest YAML parser in the ecosystem
• 🥇 Fastest TOML parser available
• 🥇 First high-performance TOON implementation
• 🥈 Top-tier JSON serialization performance
• 🥈 Competitive CSV with WASM acceleration
• 🥈 XML performance with unique dirty-cleaning
• 🥈 MessagePack with full type support

• ⚡ Blazing Fast - Up to 415x faster than alternatives
• 🛡️ Type Safety - Full TypeScript support with strict types
• 🔧 Dirty Input - Handle malformed JSON/XML automatically
• 🦀 WASM Acceleration - Rust-powered for performance-critical paths
• 📦 Zero Dependencies - Minimal bundle size
• 🔄 Streaming API - Process large files efficiently

• Dirty JSON parsing (unquoted keys, trailing commas, comments)
• Type preservation (Date, BigInt, Map, Set, RegExp, etc.)
• Schema validation (Zod, JSON Schema)
• Streaming for large files

• Full YAML 1.2 spec support
• Anchors and aliases
• Multi-document parsing
• Custom tags
• 2-415x faster than competitors

• Tables and nested tables
• Array of tables
• Inline tables
• Type-safe parsing
• 2-9x faster than alternatives

• Section support ([section])
• Comment support (; and #)
• Type coercion (numbers, booleans)
• Git/PHP/Windows INI compatible
• 2-3x faster than alternatives

• Automatic type detection
• Custom delimiters
• Header handling
• WASM acceleration
• Streaming support

• DOM and object conversion
• CDATA support
• Namespace handling
• Dirty XML cleaning (unique!)
• Matches fastest parsers

• Full MessagePack spec support
• Binary data handling
• Date/timestamp encoding
• BigInt support
• 20-50% smaller than JSON

• 30-60% token savings for LLMs
• Table format for uniform arrays
• Minimal quoting
• YAML-style indentation
• Perfect for ChatGPT/Claude/GPT-4 prompts

molt/
├── packages/
│   ├── json/          # JSON transformer
│   ├── yaml/          # YAML parser/serializer
│   ├── toml/          # TOML parser/serializer
│   ├── csv/           # CSV parser/serializer
│   ├── xml/           # XML parser/serializer
│   └── molt/          # Meta package
├── docs/              # VitePress documentation
├── .github/           # CI/CD workflows
├── .changeset/        # Changesets for versioning
├── turbo.json         # Turborepo configuration
└── BENCHMARKS.md      # Performance benchmarks

• Bun >= 1.0
• Turbo (installed via deps)

# Clone the repository
git clone https://github.com/sylphx/molt.git
cd molt

# Install dependencies
bun install

# Build all packages (with Turbo cache)
turbo build

# Run tests for all packages
turbo test

# Run benchmarks
turbo bench

# Lint and format
bun lint
bun format

# Create a changeset (for versioning)
bunx changeset

# Version packages
bunx changeset version

# Publish to npm
bunx changeset publish

# Start docs dev server
cd docs
bun dev

# Build docs
bun docs:build

# Preview built docs
bun docs:preview

• 📖 Full Documentation (coming soon)
• 📊 Benchmarks
• 🔧 API Reference
• 📦 Package Guides

• Runtime: Bun - Ultra-fast JavaScript runtime
• Language: TypeScript with strict mode
• Monorepo: Turborepo - High-performance build system
• Versioning: Changesets - Version management
• Testing: Vitest - Fast unit testing
• Benchmarking: Vitest benchmark mode
• Bundling: tsup - TypeScript bundler
• Linting: Biome - Fast linter and formatter
• Docs: VitePress - Documentation site
• Acceleration: Rust + WASM for performance-critical paths

We welcome contributions! Here's how to get started:

1. Fork the repository
2. Create a feature branch: git checkout -b feature/your-feature
3. Make your changes with tests
4. Run tests: turbo test
5. Create a changeset: bunx changeset
6. Lint: bun lint:fix
7. Commit: git commit -m "feat(json): add awesome feature"
8. Push and create a Pull Request

See CONTRIBUTING.md for detailed guidelines.

MIT © Sylphx

Each package in this monorepo is licensed under the MIT License.

Built with inspiration from:

• dirty-json - Original dirty JSON parser
• superjson - Type-preserving JSON
• js-yaml - YAML parser
• fast-xml-parser - XML parser
• papaparse - CSV parser

Made faster, more powerful, and production-ready by Sylphx.

If you find Molt useful, please consider giving it a star ⭐

Built with ❤️ by Sylphx | Website | X