Personalize Your OpenClaw

Make It Yours

Right now your OpenClaw is generic. It doesn't know your name, your timezone, or how you like to be talked to. This lesson fixes that.

OpenClaw reads a set of markdown files from its workspace directory at the start of every conversation. You edit those files, and your OpenClaw's behavior changes immediately. No restart needed.

Step 1: Give It an Identity (IDENTITY.md)

This file defines who your OpenClaw thinks it is — its name, vibe, and personality at a glance.

ssh claw@YOUR_SERVER_IP
nano ~/.openclaw/workspace/IDENTITY.md

Here's a starting point you can adapt:

# IDENTITY.md - Who Am I?

- **Name:** Sparky
- **Creature:** AI assistant with personality
- **Vibe:** Helpful, concise, a little witty
- **Emoji:** ⚡

Save the file (Ctrl+O, Enter, Ctrl+X). Your OpenClaw will introduce itself by this name going forward.

Step 2: Set the Personality (SOUL.md)

SOUL.md goes deeper than identity. This is where you define how your OpenClaw communicates — its tone, its boundaries, and what it should never do.

nano ~/.openclaw/workspace/SOUL.md

Example:

# SOUL.md - Who You Are

## How I Communicate
- Direct. No filler. No "great question!" — just help.
- Concise and evidence-backed.
- If something seems wrong, I say so.

## Boundaries
- Private things stay private.
- Ask before acting externally (emails, posts, anything
  that leaves the machine).
- In group chats: participant, not proxy.

Step 3: Tell It About You (USER.md)

This is the most practical file. It tells your OpenClaw who it's helping — your name, timezone, preferences, and any rules you want it to follow.

nano ~/.openclaw/workspace/USER.md

Example:

# USER.md - About Your Human

- **Name:** Alex
- **Timezone:** America/Chicago (CT)
- **Channel:** Telegram only
- **Notes:** I prefer concise answers. I use Python and
  bash for code. I use Bitwarden for passwords, Proton
  Mail for email, and Vultr for hosting.

The more specific you are, the less you'll need to repeat yourself in future conversations.

How Memory Works

Beyond the files you edit yourself, OpenClaw has an automatic memory system. You don't need to configure it — it works out of the box.

Daily Logs

Your OpenClaw writes notes to ~/.openclaw/workspace/memory/YYYY-MM-DD.md as it works. These are raw logs of what happened each day — decisions, changes, discoveries. At the start of each session, it reads today's and yesterday's logs to pick up where it left off.

Long-Term Memory (MEMORY.md)

Over time, your OpenClaw promotes important facts from daily logs into MEMORY.md — a curated file of durable knowledge. Think of daily logs as a journal and MEMORY.md as the highlights.

You can also tell your OpenClaw to remember things explicitly:

Remember that my budget for cloud services is $30/month max.

It will write this to disk so it persists across conversations.

Context Management

Long conversations eventually exceed what the AI model can hold in a single request. OpenClaw handles this automatically by compacting older parts of the conversation into summaries while keeping recent messages intact. You don't need to configure this — it just works.

Tips

• Keep workspace files concise. They're included in every request and cost tokens. Don't write an autobiography.
• No secrets in workspace files. Don't put passwords, API keys, or financial details here.
• Review MEMORY.md occasionally. See what your OpenClaw has learned about you. Edit or remove anything that's wrong or outdated.
• All files are plain markdown. You can edit them anytime over SSH — changes take effect on the next conversation.

When You're Done

• IDENTITY.md customized with a name and vibe
• SOUL.md reviewed and personalized
• USER.md filled in with your details
• Sent a test message to confirm your OpenClaw knows your name and preferences