ZONFORMAT – Zero overhead notation Token Reducer

Summary: ZON is a human-readable, token-efficient data format designed for large language models, reducing token usage by 35-70% compared to JSON, YAML, and TOON. It lowers API costs and integrates with tools like Langchain, llama.cpp, and OpenAI SDK within a lightweight package.

What it does

ZON compresses data into fewer tokens without parsing overhead, enabling more efficient LLM prompts and API usage. It supports streaming and schema validation with under 2KB size.

Who it's for

Developers and users working with LLMs who want to reduce token consumption and API costs.

Why it matters

It addresses high token usage in JSON-based LLM prompts, significantly cutting API expenses.