OpenClaw Setup Guide

Getting Started

What to do after you've installed OpenClaw

What We're Assuming

This guide assumes you already have OpenClaw installed, connected to a messaging channel, and running. If you haven't done that yet, check the OpenClaw documentation to get set up first.

Once you're past the onboarding flow and can message your agent, come back here.

What's Next

At this point you have a working agent, but it's a blank slate. It doesn't know who you are, it has no memory between sessions, and it's not doing anything proactively. This guide is about turning it into something actually useful.

Don't try to set everything up at once though. I built this over weeks. Here's how I'd do it if I were starting fresh.

Phase 1 -- Identity and Memory (Day 1)

The foundation. Without these, your agent has no continuity between sessions and no idea who you are.

  1. Agent Identity -- give it a name, personality, and operating instructions
  2. User Profile -- tell it who you are, your goals, how you work
  3. Memory System -- set up daily logs and long-term memory so it has continuity across sessions

About 30 minutes of prompting and iterating.

Phase 2 -- Life OS (Day 2)

This is where you start getting daily value. Set up a couple cron jobs and your agent starts showing up for you every morning.

About 20 minutes to set up. After this, your agent is doing something useful every day.

Phase 3 -- Knowledge Management (Week 1)

This is where it gets really interesting.

  • Syncthing -- if your agent runs on a different machine than your vault, set up Syncthing to sync files between them
  • LogSeq/Obsidian Vault -- your personal knowledge base, zettelkasten process, research and writing workflows
  • QMD Search -- semantic search across everything

About an hour for initial setup (plus Syncthing if you need it), then it grows as you feed it sources.

Phase 4 -- Content and Development (Week 2+)

Power-user stuff. Each of these is independent -- pick what's relevant to you.

On this page