OpenClaw Operator — Quick Start

This is the fastest way to get the repo running as a self-hosted operator workspace. The recommended first run is the root local-dev path, which starts the orchestrator and serves the canonical operator console at /operator.

Deployment Paths

Path A: Recommended First Run (Root Local Dev)

cd openclaw-operator
npm install
cp orchestrator/.env.example orchestrator/.env
# fill in the required vars in orchestrator/.env
npm run dev

Open http://127.0.0.1:3000/operator.

Path A now uses the repo-local ./orchestrator/data/orchestrator-state.json state target by default, so a first boot does not require Mongo or Redis.

Config: ./orchestrator_config.json (repo-relative paths for local dev)

Systemd service: ./systemd/orchestrator.service

If you want the always-on user-service path on 3312 after the first local boot:

mkdir -p ~/.config/systemd/user
install -m 0644 systemd/orchestrator.service ~/.config/systemd/user/orchestrator.service
systemctl --user daemon-reload
systemctl --user enable --now orchestrator

Then open http://127.0.0.1:3312/operator. The tracked service assumes the repo lives at ~/openclaw-operator; if you clone elsewhere, update the unit paths before enabling it.

---

Path B: Official Docker Demo Stack

This is the supported public Docker path. It brings up the orchestrator, MongoDB, and Redis from the repo root and serves the operator console on a localhost-only port.

cd openclaw-operator
docker compose up -d --build
npm run docker:demo:smoke

Open http://127.0.0.1:4300/operator.

Demo bearer keys:

• viewer: demo-viewer-key-local-only
• operator: demo-operator-key-local-only
• admin: demo-admin-key-local-only

The Docker quickstart uses ./orchestrator/orchestrator_config.json inside the container image and keeps the default host exposure on 127.0.0.1:4300 so it does not collide with the common repo-native dev port.

If you want real provider keys, different ports, or non-demo credentials:

cp docker-compose.override.example.yml docker-compose.override.yml
# edit docker-compose.override.yml
docker compose up -d --build
npm run docker:demo:smoke

Path C: Advanced Observability Stack

Use ./orchestrator/docker-compose.yml only when you intentionally want the heavier stack with Prometheus, Grafana, and Alertmanager.

cd openclaw-operator/orchestrator
cp .env.example .env
# fill in all required vars (see below)
docker compose up -d --build

---

Environment Variables

Local root dev vars live in ./orchestrator/.env.

Variable	Required	Notes
API_KEY_ROTATION or API_KEY	✅	Bearer auth for protected operator APIs
WEBHOOK_SECRET	✅	Security posture check — orchestrator refuses to start without it
DATABASE_URL	Optional	Enables Mongo-backed historical persistence and export surfaces
REDIS_URL	Optional	Enables Redis-backed coordination and response caching
OPENAI_API_KEY	Usually	Needed for the common agent mix
ANTHROPIC_API_KEY	Optional	Needed only for Anthropic-backed paths you enable
SLACK_ERROR_WEBHOOK	Optional	Alert delivery to Slack

Docker demo note:

• the official root Docker path ships demo-local auth/database/cache credentials directly in docker-compose.yml
• that makes first boot easy, but those values are only for localhost try-outs
• replace them through docker-compose.override.yml before any shared or non-local deployment

---

Verify OpenClaw Operator Is Running

curl http://127.0.0.1:3000/health
curl http://127.0.0.1:3000/api/knowledge/summary
curl http://127.0.0.1:3312/health
curl http://127.0.0.1:4300/health

Then open http://127.0.0.1:3000/operator for root local dev, http://127.0.0.1:3312/operator for the user service, or http://127.0.0.1:4300/operator for the Docker demo and confirm the console loads.

---

Scheduled Tasks (in-process)

Schedule	Task	What
0 23 * * * (11pm UTC)	nightly-batch	Collect leads, mark high-confidence, create digest
0 6 * * * (6am UTC)	send-digest	Send digest notification
Every 5 min	heartbeat	Health check

---

Public Proof Surface

The public proof routes are now served directly by the orchestrator. No separate proof app or ingest pipeline is required for the active local runtime.

Public read-only surfaces:

• GET /api/command-center/overview
• GET /api/command-center/control
• GET /api/command-center/demand
• GET /api/command-center/demand-live
• GET /api/milestones/latest
• GET /api/milestones/dead-letter

---

Troubleshooting

Operator won't start — for root local dev, check orchestrator/.env for API_KEY_ROTATION or API_KEY, WEBHOOK_SECRET, and your model-provider key. If you intentionally enabled Mongo or Redis, also verify DATABASE_URL and REDIS_URL. For Docker demo, run docker compose logs -f orchestrator.

Public proof looks stale — check /api/milestones/latest, /api/milestones/dead-letter, and /api/command-center/overview before assuming a queue or UI problem.

Slack alerts not arriving — verify SLACK_ERROR_WEBHOOK is set and test with:

curl -X POST "$SLACK_ERROR_WEBHOOK" -d '{"text":"test"}'

---

Deployment Checklist

• [ ] .env created with the required auth vars for Path A, or full service vars for Path C
• [ ] npm install run in the repo root (root postinstall installs orchestrator and operator console deps)
• [ ] npm run dev rebuilds the operator console bundle and starts the orchestrator from the repo root
• [ ] http://127.0.0.1:3000/operator, http://127.0.0.1:3312/operator, or http://127.0.0.1:4300/operator loads the canonical operator console
• [ ] /operator loads the built operator console instead of a fixture bundle
• [ ] Public proof routes respond from the orchestrator: /api/milestones/latest and /api/command-center/overview

Root Command Hub

The root package.json is now the default command hub for the active control plane:

cd openclaw-operator
npm run build
npm run typecheck
npm run test
npm run test:unit
npm run test:integration
npm run docs:drift
npm run docs:links
npm run clean:dry-run

npm run clean is intentionally conservative: it removes cache, coverage, and TypeScript build-info artifacts only. Frontend dist/ cleanup is opt-in via npm run clean:builds.