When troubleshooting hasn't worked and you need a clean slate, here's how to factory reset OpenClaw — on both native installs and Docker containers.

When to Factory Reset

A factory reset is the right move when:

openclaw doctor --fix didn't solve your issue
The dashboard loads but nothing works
OpenClaw doesn't respond to any commands
You see persistent connection or token errors after multiple fix attempts
You want to start over with a clean configuration

Warning: A factory reset removes your configuration, paired devices, and memory. Your AI provider API key and Telegram bot token will need to be re-entered.

Native Install Reset

Step 1: Stop OpenClaw

openclaw gateway stop

If that doesn't work:

# Find and kill the process
pkill -f openclaw

Step 2: Back Up Important Data (Optional)

# Save your current config
cp ~/.openclaw/openclaw.json ~/openclaw-backup.json

# Save memory data if you want to keep it
cp -r ~/.openclaw/memory ~/openclaw-memory-backup

Step 3: Remove the Data Directory

rm -rf ~/.openclaw

This removes:

Configuration (openclaw.json)
Device pairings
Gateway tokens
Memory data
Logs

Step 4: Reinitialize

openclaw setup

This walks you through the initial setup wizard: gateway mode, AI provider, and API key.

Step 5: Reconfigure

# Set gateway mode
openclaw config set gateway.mode local

# Set your AI provider
openclaw config set ai.apiKey "sk-ant-..."

# Reconnect Telegram (if used)
openclaw config set telegram.token "your-bot-token"

# Start
openclaw restart

Step 6: Re-Pair Devices

Open your browser or Telegram bot and approve the new device pairing:

openclaw devices list
openclaw devices approve DEVICE_ID

Docker Container Reset

Option 1: Rebuild the Container

# Stop and remove the container
docker stop YOUR_CONTAINER_ID
docker rm YOUR_CONTAINER_ID

# Remove the data volume (this is the actual reset)
docker volume rm openclaw-data

# Recreate
docker run -d \
  --name openclaw \
  -v openclaw-data:/app/data \
  -p 3001:3001 \
  -e OPENCLAW_GATEWAY_TOKEN="new-stable-token" \
  -e ANTHROPIC_API_KEY="sk-ant-..." \
  openclaw-stack

Option 2: Reset Inside the Container

# Enter the container
docker exec -it YOUR_CONTAINER_ID bash

# Remove the data directory
rm -rf /app/data/*

# Re-run setup
openclaw setup

# Exit
exit

# Restart the container
docker restart YOUR_CONTAINER_ID

Option 3: Docker Compose Reset

# Stop everything
docker-compose down

# Remove volumes
docker-compose down -v

# Rebuild and start
docker-compose up -d

Partial Reset Options

You don't always need a full factory reset. Here are targeted resets:

Reset Config Only

openclaw config reset
openclaw config set gateway.mode local
openclaw config set ai.apiKey "your-key"

Reset Devices Only

# Remove all paired devices
openclaw devices list
openclaw devices remove --all

Reset Gateway Token Only

openclaw config unset gateway.token
openclaw doctor --generate-gateway-token
openclaw restart

Clear Memory Only

rm -rf ~/.openclaw/memory
openclaw restart

After the Reset

Post-reset checklist:

Verify gateway starts: openclaw status should show gateway: running
Check logs: openclaw logs --tail 20 — look for [gateway] listening on
Run doctor: openclaw doctor — should show all checks passing
Test Telegram: Send a message to your bot
Test browser: Open the dashboard URL

When Reset Doesn't Help

If a factory reset doesn't fix your issue, the problem is likely outside OpenClaw:

Network/firewall blocking the gateway port
Node.js version too old (need 22+)
Disk full — no space for data files
OS-level permission issues

Check:

node --version        # Need 22+
df -h                 # Check disk space
openclaw logs --tail 50  # Look for OS-level errors

Skip the Reset Cycle

ClawTank manages your OpenClaw instance with automated health checks and recovery. If something goes wrong, use the dashboard to rebuild your container in one click — no manual reset needed.