id-agent: Token-Efficient Word-Based IDs for AI Agents Reduce LLM Token Costs

Summary

id-agent is a new open-source library that generates token-efficient, word-based IDs for AI agents, designed to reduce token consumption in LLM context windows. While UUIDs cost ~23 tokens and are prone to hallucination by LLMs, id-agent produces memorable word-based IDs at ~14 tokens (default 8-word format) with equivalent collision resistance. The library offers configurable word counts (5-8 words) for different security needs, making it the first ID library built specifically for the context window rather than the database.