Compress conversation context deterministically to reduce token usage.
Token compression for AI contexts, reducing token consumption by compressing conversation exchanges before they enter the LLM context window.
