Claw Compactor is a token-compression engine for LLM agents: a 14-stage, content-aware pipeline shrinks code, JSON, logs and diffs by 15-82%, with reversible markers so agents can retrieve the original context.
2005ai-agent-tools,context-compression,context-pruningClaw Compactor is a token-compression engine for LLM agents: a 14-stage, content-aware pipeline shrinks code, JSON, logs and diffs by 15-82%, with reversible markers so agents can retrieve the original context.