← Back to OpenClaw guide
Category 1

Setup and basics

Everything a new operator needs to get OpenClaw working, believable, and useful without getting lost in advanced branches too early.

Use this whenYou are installing, onboarding, learning the first commands, or fixing the most common early confusion.
Start here

The shortest safe path to a working setup

For most people the correct order is simple: install OpenClaw, complete onboarding, confirm the core commands work, then try one useful workflow before touching deeper config. Most early frustration comes from skipping that order and diving into advanced knobs too soon.

  1. Install cleanly. Use the standard install first and avoid speculative source work.
  2. Run onboarding fully. Get one believable session working before chasing edge cases.
  3. Verify the basics. Use the small command set that proves the runtime is healthy.
  4. Try one real workflow. Make the system do something concrete so confidence comes from results.
Install OpenClawStart from the normal install path.First setup and onboardingTurn a fresh install into a real first session.Everyday commands that matterUse the small set that actually proves health.
What to do right away

First workflows, first prompts, first recovery moves

First useful workflow

Ask OpenClaw to inspect a real file, verify a live status, or summarize a concrete artifact. The first win should feel practical, not theatrical.

First prompts that work

Good starter prompts are specific, bounded, and verifiable. Ask for inspections, checks, and edits with a clear success condition.

First confusion points

If something feels wrong early on, check whether the issue is config, browser policy, missing permissions, or simply a vague ask.

First useful workflowsGet to useful outcomes fast.What to ask OpenClaw firstUse stronger starter prompts.Common first-run questionsResolve the normal early friction points.
Config essentials

What actually matters before advanced tuning

Most users do not need a sprawling config study session. They need to know which model path is active, how to edit config safely, and how to recover from malformed JSON or risky browser/network settings without making things worse.

Config overviewGet the whole config map.How config worksSee how the main config surface is organized.Safe editingEdit carefully and avoid self-inflicted breakage.TroubleshootingRecover from bad edits and validation problems.
Open the deeper reference links for this category
Schema, patching, and validationFull getting started laneFull config lane