OpenClaw

A self-hosted personal AI assistant that can browse the web, run commands, and connect to chat platforms like Slack, Discord, and Telegram.

Coding Agents

Versionlatest

LicenseMIT License

InterfaceWeb UI

GPUOptional

One-Click Install on Laboratory OS

GitHub

Securely deploy OpenClaw on Laboratory OS.

OpenClaw

Laboratory OS Installation Details

Environment

Working Directory/workspace/openclaw

Environment Variables

APP_DIRECTORY

/workspace/openclaw

PORT

OPENCLAW_PROVIDER_URLBase URL for the model provider API (e.g., http://localhost:8000/v1)

—

OPENCLAW_API_KEYAPI key/token for the model provider. Leave blank for local providers that do not require authentication.

—

OPENCLAW_MODEL_IDModel identifier to use with the provider (e.g., llama-3.3-70b, gpt-4o)

—

OPENCLAW_API_TYPEAPI compatibility type. Use "openai-completions" for OpenAI-compatible endpoints or "anthropic-messages" for Anthropic-compatible endpoints.

openai-completions

OPENCLAW_GATEWAY_TOKENToken to secure access to the OpenClaw gateway API. If not set, a random token will be generated.

—

Install & Run

install-and-run.sh

#!/bin/bash
set -e

export PATH="/opt/conda/bin:/root/.local/bin:$PATH"

# Create the directory if it doesn't exist
mkdir -p "/workspace/openclaw"
  
# Change to the app directory
cd "/workspace/openclaw"


# Environment setup complete
echo "Environment setup complete!"

npm install -g openclaw@latest

# Configure gateway settings
openclaw config set gateway.mode "local"
openclaw config set gateway.auth.mode "token"
GATEWAY_TOKEN="${OPENCLAW_GATEWAY_TOKEN:-openclaw-local}"
openclaw config set gateway.auth.token "$GATEWAY_TOKEN"
openclaw config set gateway.remote.token "$GATEWAY_TOKEN"
openclaw config set gateway.trustedProxies '["127.0.0.1"]'
openclaw config set gateway.bind "lan"
openclaw config set gateway.port "$PORT"
openclaw config set gateway.controlUi.allowedOrigins '["*"]'

# Configure custom provider if all required vars are set
if [ -n "$OPENCLAW_PROVIDER_URL" ] && [ -n "$OPENCLAW_MODEL_ID" ]; then
  API_TYPE="${OPENCLAW_API_TYPE:-openai-completions}"
  API_KEY_VALUE="${OPENCLAW_API_KEY:-local}"

  cat > /tmp/openclaw-provider.json <<PROVIDER_EOF
{"baseUrl": "${OPENCLAW_PROVIDER_URL}", "apiKey": "${API_KEY_VALUE}", "api": "${API_TYPE}", "models": [{"id": "${OPENCLAW_MODEL_ID}", "name": "Custom Model"}]}
PROVIDER_EOF
  openclaw config set models.providers.custom "$(cat /tmp/openclaw-provider.json)"
  openclaw config set agents.defaults.model.primary "custom/${OPENCLAW_MODEL_ID}"
  rm -f /tmp/openclaw-provider.json
fi

# Background process to auto-approve device requests for out-of-the-box usability
(
  set +e
  sleep 5
  while true; do
    PENDING=$(openclaw devices list 2>&1 | grep -oE "[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}" | head -1)
    if [ -n "$PENDING" ]; then
      openclaw devices approve "$PENDING" 2>&1 || true
    fi
    sleep 2
  done
) &
APPROVER_PID=$!

# Start gateway
openclaw gateway $OPENLAB_EXTRA_ARGS &
GATEWAY_PID=$!

# Wait for gateway process
wait $GATEWAY_PID
kill $APPROVER_PID 2>/dev/null || true

CLI Command

cli.sh

# OpenClaw
exec openclaw