token-optimizer

Overview: Slash your OpenClaw AI costs from $1,500+/month to under $50/month with the Token Optimizer skill. This skill reduces costs by 97% through model routing, free Ollama heartbeats, prompt caching, and budget controls. Key Features: * Model routing (Haiku by default, Sonnet/Opus only when needed) * Multi-provider heartbeats (Ollama, LM Studio, Groq, or disable entirely) * Session management (load 8KB instead of 50KB context) * Caching (reuse prompts at 10% cost) * New features: rollback, health check, diff preview, and CI/pipeline friendly output How It Works: The Token Optimizer skill modifies the OpenClaw configuration to optimize costs. It uses a one-command optimizer with diff preview, multi-provider heartbeat, config rollback and health check commands, and ready-to-use config templates. Use Cases: * Reduce AI costs from $1,500/month to $50/month * Optimize model routing for Haiku, Sonnet, and Opus * Use free Ollama heartbeats instead of paid API heartbeats * Cache prompts to reduce costs by 90% * Use rollback and health check commands for quick system status and config backup management