Myway + OpenClaw Setup

Myway is a self-hosted personal AI home screen — a local-first PWA that gives you a phone-style dashboard powered by AI. This skill sets it up using OpenClaw as the AI backend.

When to use this skill

• User wants to install Myway and already has OpenClaw running locally
• User asks about setting up a personal AI dashboard with OpenClaw
• User wants to configure or reconfigure their Myway + OpenClaw connection

Prerequisites

Check these before starting:

1. Node.js 22+ — run node -v to verify (must show v22 or higher)
2. yarn — run yarn -v to verify. If missing: npm install -g yarn
3. OpenClaw — run openclaw gateway status or check if ~/.openclaw/openclaw.json exists

If prerequisites fail, stop and help the user install them first.

Installation

Optionally ask the user for:

1. Their OpenClaw gateway URL (default: http://localhost:18789)
2. A gateway token (usually not needed for localhost)
3. A directory for the file browser (default: ~/vault)
4. Where to install (directory name, default: myway)

Then run the setup:

npx @uchibeke/myway <directory> \
  --ai-mode openclaw \
  --openclaw-url http://localhost:18789 \
  --root ~/vault \
  --port 48291 \
  --start

All flags are optional except --ai-mode openclaw. Defaults:

• directory: myway
• --openclaw-url: http://localhost:18789
• --openclaw-token: empty (not needed for localhost)
• --root: ~/vault
• --port: 48291

Add --no-aport to skip APort guardrails setup.

The command will:

1. Scaffold a new directory with the Myway source
2. Install dependencies (2-5 minutes)
3. Write .env.local with the provided config
4. Initialize the database
5. Build for production (1-2 minutes)
6. Start the server and open the browser (if --start is passed)

Important: This command takes 3-7 minutes to complete. Do not interrupt it.

Verify it worked

After the command completes with --start, the server should be running. Verify:

curl -s -o /dev/null -w "%{http_code}" http://localhost:48291/

Should return 200. If the user didn't pass --start:

cd <directory>
yarn start

Background running with PM2

For long-running deployments:

cd <directory>
cp ecosystem.config.cjs.example ecosystem.config.cjs
pm2 start ecosystem.config.cjs

Reconfiguring

To reconfigure an existing install:

cd <existing-myway-directory>
npx @uchibeke/myway --setup \
  --ai-mode openclaw \
  --openclaw-url http://localhost:18789

Or edit .env.local directly, then rebuild:

cd <existing-myway-directory>
# Edit .env.local with new values
yarn build
# Restart: yarn start, or pm2 restart myway

OpenClaw connection details

Myway connects to OpenClaw via its gateway API:

• Default URL: http://localhost:18789
• Config file: ~/.openclaw/openclaw.json (auto-detected at runtime)
• Auth: Optional gateway token, not needed for localhost setups
• How it works: Myway sends chat/skill requests to OpenClaw's gateway, which routes them to the configured AI model

Troubleshooting

Problem	Fix
OpenClaw not detected	Run openclaw gateway status to verify it's running
Connection refused	Check OPENCLAW_BASE_URL in .env.local matches openclaw gateway status output
Port already in use	Change PORT in .env.local or: fuser -k 48291/tcp
Black screen after restart	Run pm2 logs myway --err. Usually EADDRINUSE — kill stale process and restart PM2
Build fails	Ensure Node.js 22+ and run yarn install first
Command hangs	Normal — dependency install takes 2-5 min, build takes 1-2 min

Environment variables

All config lives in <directory>/.env.local. Key variables for OpenClaw mode:

PORT=48291
MYWAY_ROOT=~/vault
OPENCLAW_BASE_URL=http://localhost:18789
# OPENCLAW_GATEWAY_TOKEN=  # only if gateway requires auth
MYWAY_SECRET=<auto-generated>

See .env.local.example in the install directory for all options (Google OAuth, Telegram, TTS, etc.).