OpenClaw Command Center
Monitor OpenClaw sessions, costs, system health, and scheduled jobs from one local dashboard.
- Source repo
- jontsai/openclaw-command-center
- Stars
- ★ 250
- Last updated
- 1mo ago
- License
- MIT
- Primary language
- JavaScript
- FA score
- 62/100 · Some gaps
At a glance
- How it runs
- Works with
- Platform-specific
- Cost
- Free, no paid service needed
- Setup effort
- Low · running in minutes
- You'll need
- Typical use
- An OpenClaw deployment maintainer who wants one view of active sessions, models, channels, and usage.
- Not a fit if
- Teams monitoring agents outside an OpenClaw workspace
- Teams that need a graphical chat composer for human-agent interaction
- Website
- clawhub.ai/jontsai/command-center ↗ set on GitHub
- Maintenance
- Commits in the last 6 monthslast commit 2026-10-06latest release v1.4.1 (2026-03-08)
- Source review
- 62/100 · Some gaps
What does this agent do, and when should you use it?
OpenClaw Command Center is a self-hosted monitoring dashboard for OpenClaw deployments, built from a Node.js HTTP server and a browser UI. The server exposes a unified state REST API and an SSE stream for live updates; the UI receives updates about every two seconds, with a five-second cache. It reads deployment data such as sessions, usage, jobs, and memory from an OpenClaw workspace, and can also report system vitals; Cron Jobs are retrieved through OpenClaw's read-only CLI API. It binds to 127.0.0.1 by default and supports token, Tailscale, Cloudflare Access, or IP allowlist authentication. It has no chat composer; the optional Spacesuit extension adds per-agent flavors, domain panels, and workflows.
On startup, the server auto-detects an OpenClaw workspace, or accepts its root through OPENCLAW_WORKSPACE, and reads deployment data for dashboard views such as sessions, usage, and health. GET /api/state returns unified dashboard state; GET /api/health, GET /api/vitals, and GET /api/sessions provide dedicated endpoints; GET /api/events streams live updates over SSE. The UI presents sessions, token usage and costs, system vitals, Cron Jobs, Cerebro topics, operators, and memory files. Cron Jobs are listed through OpenClaw's read-only CLI API and cached for 30 seconds. Users can enter a monthly spend baseline and compare it with a 30-day projection based on the selected calendar-day average; this is a what-if comparison, not realized savings or a billing forecast.
- An OpenClaw deployment maintainer who wants one view of active sessions, models, channels, and usage.
- An operator checking CPU, memory, disk, and temperature metrics from a local dashboard.
- An automation maintainer who needs to inspect OpenClaw Cron Jobs and their status.
- A team comparing estimated per-model costs with a custom monthly spend baseline.
- An operator who needs to hide sensitive topics, sessions, or jobs for a demo or screenshot.
Key setup facts
Pulled from the install and usage notes below. The project README has the current values.
- Install / run
npx clawhub@latest install command-center
- Ports
3333- Environment variables
OPENCLAW_PROFILEDASHBOARD_TOKEN- Runtime
- Node.js 18
- Install notes cover
- macOS
How do you install or deploy this agent?
Requires Node.js 18 or newer. The README documents installation through ClawHub and running from a direct Git clone. ClawHub path:
npx clawhub@latest install command-center
cd skills/command-center
node lib/server.jsDirect clone path:
git clone https://github.com/jontsai/openclaw-command-center
cd openclaw-command-center
node lib/server.jsThe dashboard runs at http://localhost:3333 by default. Some system-vitals features optionally use sysstat, lm-sensors, or macOS temperature tools; without them the dashboard still runs, while affected metrics may show zeros or fallback values.
How do you use this agent?
Open http://localhost:3333 in a browser. The server searches OPENCLAW_WORKSPACE, common workspace paths, and common directory names; OPENCLAW_WORKSPACE and OPENCLAW_PROFILE can also be configured. Multi-profile examples:
node lib/server.js --profile production --port 3333
node lib/server.js --profile dev --port 3334Choose an authentication mode as needed, for example:
DASHBOARD_AUTH_MODE=token DASHBOARD_TOKEN=secret node lib/server.jsThe main endpoints are GET /api/state, GET /api/health, GET /api/vitals, GET /api/sessions, and GET /api/events. To run a minimal core-only view:
COMMAND_CENTER_MODE=core npm startWhat are this agent's strengths and limitations?
- GET /api/state consolidates dashboard data into one endpoint, while GET /api/events provides SSE live updates.
- It binds to 127.0.0.1 by default and offers token, Tailscale, Cloudflare Access, and IP allowlist authentication modes.
- It can auto-detect an OpenClaw workspace, and Spacesuit is not required for the core dashboard.
- It brings session, system-vitals, scheduled-job, cost, and memory views into one operations interface.
- Its documented purpose depends on an OpenClaw workspace; support for monitoring other agent platforms is not established.
- The README says it has no graphical human-agent interaction surface or chat composer.
- Some system metrics rely on optional OS tools; without them, metrics may be zero or use fallback values.
- Cost and savings figures are estimates or what-if comparisons, not actual billing or realized savings.
How does this agent compare with similar options?
The README positions Command Center as a standalone OpenClaw monitoring dashboard and Spacesuit as a recommended optional extension: Command Center owns the common dashboard and supported controls, while Spacesuit adds per-agent flavors, domain panels, and workflows.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| OpenClaw Command Center This agent | 62 · Some gaps | Agent plugin / skillFree | ★ 250 | 1mo ago | JavaScript | — |
| ClawMetry | 71 · Some gaps | CLIFreemium | ★ 439 | 1d ago | Python | Codex · Claude Code |
| OpenClaw Monitor | 36 · Major gaps | Self-hosted serviceFree | ★ 13 | 6mo ago | JavaScript | — |
| AgentOps | 46 · Major gaps | Library / SDKFreemium | ★ 6k | 4mo ago | Python | OpenAI API · Claude API |
How does FollowAgents rate this agent?
Why each dimension lost points
The source describes a monitoring dashboard, a localhost default, read-only defaults, authentication modes, no telemetry, and no secret display. Optional extensions and project snapshots are described as read-only, supporting relatively strong least-privilege, data-flow, and external-effects scores. Authentication tests cover tokens, proxy identity headers, and IP allowlists, but the supplied server source is incomplete and does not establish authorization behavior for every route; evidence that all side-effecting actions require confirmation is also limited. Privacy controls, account-isolated caching, and a plan to avoid exposing account email show some sensitive-data safeguards, though the packet is partial. Dependency security earns only limited credit: a small set of development dependencies and npm ci are shown, but no audit or lockfile evidence is included. Data migration is mentioned without recovery guidance, limiting rollback. The README, license, repository metadata, and ClawHub workflow provide attribution clues, while publisher identity remains unknown as instructed.
The documentation distinguishes read-only directory access, optional AI Jobs, and core functionality. Sample tests support account-isolated caching, stale-state labeling, and unknown-data handling. However, the README's zero-user-dependencies claim sits alongside requirements for an OpenClaw workspace and system packages for some features. Cache behavior, asynchronous reads, independent errors, and unavailable states receive some documentation. Failure handling is partially supported by tests, but the sample is not broad enough to establish consistency across all error paths.
The intended audience and scenarios center on monitoring OpenClaw deployments, costs, and scheduled work, with multi-profile support, core mode, mobile layouts, and several authentication options. The docs clearly state that extensions do not provide chat and that portfolio data comes from snapshots without changing external records. Trigger conditions are mostly limited to manual startup and background refresh; the packet does not fully specify every automatic behavior or applicability condition. Node 18+, workspace auto-detection, configuration variables, and optional system dependencies give useful environment guidance, with limited evidence for other environments.
The README organizes features, installation, configuration, API, and architecture, and links to skill and architecture documentation. Installation instructions include an alternative Git path, optional dependencies, and multi-profile examples. Terms such as Cron Jobs and AI Jobs are distinguished. Feature descriptions and examples are substantial, though no FAQ is shown. The docs disclose that cost figures are estimates and describe snapshot limits and missing optional metrics. The complete MIT license text merits full marks. package.json and SKILL.md versions are synchronized and the release process is documented, but the packet does not include actual changelog entries. AGENTS.md describes PR and publishing responsibilities; it is primarily contributor guidance, and the maintenance evidence still relies partly on unverified identity claims.
The dashboard combines sessions, costs, system status, and jobs, with SSE, cost breakdowns, and privacy controls, showing clear monitoring utility. A unified state endpoint and local operation reduce user effort. The product has no graphical chat interaction, cost and savings figures are estimates, and usefulness depends on available OpenClaw data sources, so output usability, marginal value, and cost-benefit receive solid but not fully substantiated scores.
Some README claims about security, caching, read-only snapshots, and cost estimates correspond to supplied source snippets, tests, and architecture references. Tests corroborate selected authentication and directory-cache behavior, but the packet samples only part of the repository, so not every claim can be checked. The README explicitly frames savings comparisons as scenarios rather than billing forecasts, and the plan distinguishes account quotas from task costs, showing a reasonably clear separation of facts and inference. Some feature claims remain documentation-only.
- This is a low-confidence static review. The supplied source is partial and cannot establish access control for every route or confirmation behavior for every operation.
- Cost, savings, and system metrics include estimates or depend on data-source availability; do not treat them as billing forecasts or complete telemetry.
- Before deployment, check the authentication mode, proxy identity-header setup, and the scope of workspace data access.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
Release notes excerpt
* chore(deps-dev): bump minimatch from 3.1.2 to 3.1.5 (#46) (aef5b2b) * fix: SSE connection status stuck on 'Connecting...' due to i18n race (#47) (85d9039) * fix: prevent iostat process leak on macOS (#48) (f327df7) * chore: bump version to 1.4.1 (#49) (0c76755) What's Changed * chore(deps-dev): bump minimatch from 3.1.2 to 3.1.5 by @dependabot[bot] in https://github.com/jontsai/openclaw-command-center/pull/46 * fix: SSE connection status stuck on 'Connecting...' due to i18n race by @jontsai in https://github.com/jontsai/openclaw-command-center/pull/47 …View on GitHub ↗ -
Release notes excerpt
* docs: document optional system dependencies (#35) (#43) (eed7eff) * security: harden child process execution and add action allowlist (#30) (#44) (f133520) * chore: bump version to 1.4.0 (#45) (64fce27)View on GitHub ↗ -
Release notes excerpt
* chore: update SKILL.md version to 1.2.0 (2202f9e) * fix: resolve iostat resource leak and add Linux disk I/O parsing (#31) (#32) (41a60b3) * fix: ensure system paths in runCmd for LaunchAgent/cron environments (#23) (5d76bce) * fix: port temperature reading to async vitals + fix Intel Mac detection (#25) (22d71f7) * style: fix prettier formatting in server.js (8d5ec6f) * fix: parse vm_stat page size dynamically for Intel Mac compatibility (#24) (36303e4) * fix: remove dead getSystemVitalsOLD_DISABLED function (21839dd) …View on GitHub ↗
FAQ
Do I need a paid model API to run the dashboard?
Can the dashboard change agents or jobs?
Can I use it without Spacesuit?
Do cost comparisons show actual savings?
Which port does OpenClaw Command Center use?
Is OpenClaw Command Center open source?
Is OpenClaw Command Center safe to use?
Is OpenClaw Command Center still maintained?
Who makes OpenClaw Command Center?
What is OpenClaw Command Center's official website?
Thanks — we use this to improve the page.