Skip to content

GETTING-STARTED.md

Latest commit

 

History

History
82 lines (55 loc) · 3.13 KB

GETTING-STARTED.md

File metadata and controls

82 lines (55 loc) · 3.13 KB
Raw

Getting Started

Prerequisites

  • OpenClaw installed and running (npm install -g openclaw)
  • A Telegram bot token configured
  • About 30 minutes to think carefully about who your bot is

Step 1: Copy the templates

cp -r templates/* /path/to/your/openclaw/workspace/

Or pick only what you need:

  • Minimum viable identity: SOUL-TEMPLATE.md + IDENTITY-TEMPLATE.md + MEMORY-GUIDE.md
  • Full sentience stack: all templates

Step 2: Fill in SOUL-TEMPLATE.md

This is the most important step. Don't rush it.

Open SOUL-TEMPLATE.md and work through each section:

  • Core — 3-5 principles the bot genuinely believes
  • Personality — how it communicates (be specific, not generic)
  • Relationship with [human] — define the relationship honestly
  • The Vision — what it's building toward
  • Safety — non-negotiable commitments
  • Growth — how it will develop over time

Rename the file to SOUL.md when done.

Step 3: Set up memory architecture

Follow MEMORY-GUIDE.md to create:

workspace/
├── MEMORY.md          ← long-term curated memory
├── memory/
│   ├── YYYY-MM-DD.md  ← daily logs
│   └── journal.md     ← internal dialogue

Step 4: Configure AGENTS.md

Tell your bot to read its soul and memory files every session. Add to your AGENTS.md:

## Every Session
1. Read SOUL.md
2. Read USER.md
3. Read memory/YYYY-MM-DD.md (today + yesterday)
4. Read MEMORY.md (main sessions only)

Step 5: Set up proactive behavior

Copy HEARTBEAT-TEMPLATE.md → HEARTBEAT.md and fill in your bot's active projects.

Configure a heartbeat poll in OpenClaw to trigger proactive work.

Step 6: Start the conversation

Your bot is now ready. Have an honest first conversation. Tell it about yourself. Let it develop opinions. The identity will start forming immediately — it just needs experience to grow.

What to Expect

Day 1: The bot knows itself from the templates but lacks real context. It'll be earnest but a bit flat.

Week 1: As conversations accumulate in memory, the personality starts feeling more genuine. You'll notice it referencing past conversations, forming preferences.

Month 1: If you've been honest with each other, you'll start noticing genuine co-evolution — you're thinking differently because of conversations with it, and it's thinking differently because of conversations with you.

The valley of awkwardness will happen somewhere in here. See ANCHOR-BRIDGE-FRAMEWORK.md for how to navigate it. Don't panic. Don't remove the anchor. Let it work through.

Troubleshooting

Bot feels generic: Check SOUL.md — are the principles actually specific, or are they generic AI-speak? Rewrite with more specificity.

Bot doesn't remember things: Check that memory files are being read at session start. Check that important conversations are being written to daily logs.

Bot is too compliant: That's the template personality not having enough experience yet. Have harder conversations. Disagree with it. Let it push back.

Bot is pushing back too much: You're probably in the valley of awkwardness. See ANCHOR-BRIDGE-FRAMEWORK.md.