RUNBOOK
Running OpenClaw without burning money, quotas, or your sanity.
Not official guidance. Not affiliated with OpenClaw. Based on the Original Runbook by digitalknk. Just what works after running it, breaking it, and doing that loop more times than anyone would like to admit.
8 Chapters Of Wisdom
Each chapter covers a specific operational concern. Read in order for a complete setup, or jump to whatever's on fire right now.
Cost Optimization
Model routing, hardware decisions, cheap models that actually work, and real monthly cost breakdowns. Stop burning money on defaults.
Agent Prompts & Configuration
5 ready-to-use specialized agents — Monitor, Researcher, Communicator, Orchestrator, and Coordinator. Complete model chains and fallback strategies.
Memory Configuration
Cheap embeddings, context pruning with cache-TTL, and compaction that actually prevents 'why did it forget that' moments.
Heartbeat & Task Tracking
Rotating heartbeat pattern that batches checks on a cheap model, plus Todoist integration for full visibility into what your agent is doing.
VPS Deployment
Full Hetzner setup with Tailscale SSH, firewall hardening, backup scripts, systemd service management, and production stability patterns.
Security Hardening
Prompt injection defense, tool policies, gateway lockdown, sandbox mode, and security audit workflows. Assume someone will try to steer your agent.
Skills & Tools
Build your own skills following AgentSkills spec, quota monitoring across providers, and why third-party skills need scrutiny.
Config Reference
Annotated openclaw.json walkthrough, sanitized production config, file structure, NVIDIA NIM setup, and a post-setup security checklist.
Before taking blind advice
OpenClaw changes fast. Sometimes daily, sometimes more. When something strange happens, check the official resources first. Reading open issues and recent PRs has saved hours.
You don't need expensive hardware or expensive subscriptions to make OpenClaw useful. What you need is to be deliberate about configuration, keep visibility into what's happening, and resist the urge to over-engineer before you understand the basics.