toktoon 0.1.2

Lightweight token compression and JSON squashing for LLM prompts

TokToon

TokToon is a lightweight Python library designed to reduce Large Language Model (LLM) prompt token size without losing structure or intent.

It converts verbose text and JSON inputs into compact, schema-aware, and LLM-friendly formats—helping developers reduce token usage, improve latency, and lower API costs.

TokToon is deterministic, easy to integrate, and works with any LLM or agent framework.

---

Key Features

• 🔥 JSON Squasher

  • Converts nested JSON into a compact, schema-based format
  • Preserves structure while removing unnecessary syntax
  • Ideal for passing structured data to LLMs

• 🎚️ Lossless and Lossy Modes

  • Lossless mode preserves all fields and values
  • Lossy mode enables aggressive compression for maximum token savings

• 🧠 LLM-Friendly Output

  • Schema-first representation improves model reasoning
  • Avoids heavy brackets, repetition, and verbosity

• 🐍 Lightweight & Dependency-Free

  • No model or API dependencies
  • Safe for production environments

---

Example

Input JSON

{
  "friends": ["ana", "luis", "sam"],
  "hikes": [
    { "id": 1, "name": "Blue Lake Trail", "distanceKm": 7.5 },
    { "id": 2, "name": "Ridge Overlook", "distanceKm": 9.2 }
  ]
}