📦 Releases  ·  💬 Discussions  ·  🐛 Issues  ·  📘 Ask DeepWiki   |   English  ·  中文  ·  日本語

Quick Start

CLI + Web Console (recommended)

npm i -g clawmaster
clawmaster                   # start the web console

Open http://localhost:16223 — the setup wizard walks you through OpenClaw engine detection and LLM provider configuration. No config files to edit.

clawmaster serve --daemon    # run in background
clawmaster stop              # stop the service
clawmaster doctor            # verify your environment

Note

The current release is v0.3.1. The next milestone is v0.4.0 — already shipping features land there as they merge.

Desktop App (Beta)

Download the installer for your platform from GitHub Releases:

Platform	Format
macOS Apple Silicon	.dmg
macOS Intel	.dmg
Windows x64	.msi, .exe
Linux x64	.deb, .AppImage

Warning

Desktop builds are in beta. The CLI + Web Console is the recommended and most thoroughly tested install method.

From source

git clone https://github.com/openmaster-ai/clawmaster.git
cd clawmaster
npm install
npm run dev:web              # web console + backend
npm run tauri:dev            # desktop app

Requires Node.js 20+. Tauri desktop builds also need Rust — see tauri.app/start/prerequisites.

After Launch

1. Pick an existing OpenClaw profile or create a new one.
2. Connect at least one model provider and set a default model.
3. Add channels, plugins, skills, or MCP servers as needed.
4. Enable gateway or observability when you need runtime inspection.

Pick your learning path

• 🧪 Hands-on — run through clawmaster-workshop — trilingual (EN / 中文 / 日本語) tasks grouped by the six core capabilities, plus dated labs that chain tasks into real scenarios. Best if you want to do the thing.
• 🖼️ Pictured walkthrough — skim the Product Tour below. Each screenshot maps to a concrete task, so you can understand what the product does without installing anything.

Why ClawMaster

Most OpenClaw tooling stops at configuration. ClawMaster is your OpenClaw companion for real life — a bridge between OpenClaw's power and everyday usability. It's for people who want OpenClaw to actually work in their daily life (not just be correctly configured), who don't want to live in JSON and terminals, and who manage OpenClaw for a team or family.

Memory Highlights

Memory is the backbone of the Save capability. We build on PowerMem (Python · TypeScript SDK · OpenClaw plugin) instead of rolling our own:

• Native OpenClaw citizen — PowerMem already ships an OpenClaw memory plugin, so agent turns get auto-recall / auto-capture for free.
• Smart extraction, not chunk dumping — distills conversations into durable facts with Ebbinghaus-style decay and recall, which matches our "build but also maintain" direction.
• Multi-agent isolation built in — scopes per user / agent / workspace without us reinventing identity plumbing.
• Database-grade durability — pairs with OceanBase seekdb for hybrid vector + full-text + SQL, with SQLite as a cross-platform fallback.
• Open source with cross-language SDKs — we're not locked into one runtime; consistent semantics from JS to Python to Go.

Shipped

• Managed PowerMem runtime with an OpenClaw bridge across web, backend, and desktop — agent turns get auto-recall and auto-capture out of the box.
• Local workspace import that pulls markdown / memory/ into managed PowerMem, using seekdb where available and SQLite as a fallback.
• First end-to-end memory-backed skill: a daily package download digest with period-over-period deltas.
• Memory-adjacent observability — per-session spend, scheduled cost digests, and models.dev pricing.

Next (v0.4.0): full seekdb hybrid retrieval and a self-maintaining LLM Wiki module — persistent wiki pages that compound with every ingest, with Ebbinghaus decay and freshness-weighted ranking keeping content alive. See the v0.4.0 milestone for tracked work.

Product Tour

Setup wizard · 2-step guided install, tiered providers	Overview · Runtime health, next-step task flows	Models · Multi-provider config with live key validation	Channels · Guided onboarding for 6 messaging platforms
Observe · ClawProbe-backed cost, tokens, session health	Memory · PowerMem runtime with seekdb / SQLite fallback	MCP · Servers, endpoints, and skill definitions	Skills · ClawHub marketplace with install and audit

Who It Is For

• "I want OpenClaw useful in real life, not just configured." — closes the gap between install and outcome.
• "I'm non-technical but want a powerful AI assistant." — guided setup, guided usage, no JSON required.
• "I manage OpenClaw for my team or family." — one place for channels, runtime state, and onboarding.
• "I'm building advanced agent workflows." — provider management, observability, memory, sessions, plugins, skills, and MCP in one place.

Roadmap

Six core capabilities — each moves from infrastructure toward real daily use:

#	Capability	Status	What's here	What's next
1	Setup	Available	Guided wizard, 6+ LLM providers with key validation, 6 channel types (Feishu / WeChat / Discord / Slack / Telegram / WhatsApp), profile switching	One-click environment migration (#1), Windows + WSL2 first-class support
2	Observe	Available	ClawProbe-backed dashboard, per-session cost and token tracking, gateway health monitoring	Historical spend analytics, anomaly alerts, multi-profile comparison
3	Save	In progress	Managed PowerMem runtime + OpenClaw bridge, local workspace import, first memory-backed skill — see Memory Highlights	Full seekdb hybrid retrieval, self-maintaining LLM Wiki — see v0.4.0 milestone
4	Apply	In progress	PaddleOCR pipeline (upload → parse → structured markdown), layout-aware extraction	Photo → flashcard automation, invoice extraction templates, more scenario-first guided workflows
5	Build	Planned	Plugin/skill install and toggle, MCP server management, skill security auditing	Visual agent composer for skill chaining, LangChain Deep Agents integration, conversational agent builder
6	Guard	Planned	Skill Guard security scanning (dimension/severity/risk scoring), basic capability gating	API key vault (encrypted at rest), per-profile spend caps, RBAC for team deployments

Browse label:roadmap to pick up an item. Leave a comment before starting so work does not overlap.

Versioning

ClawMaster follows Pride Versioning — PROUD.DEFAULT.SHAME:

Segment	When to bump
Proud	A release you are genuinely proud of
Default	Normal, solid releases
Shame	Fixing something too embarrassing to talk about

Pre-release tags (-rc.N) are used for release candidates.

📰 News

• 2026-04-25 🚀 v0.3.0 — first official release. Setup wizard, PaddleOCR, ERNIE image, cost observability, cron management, bundled skills refresh, and managed PowerMem support. CLI is the recommended install method; desktop builds remain beta.
• 2026-04-17 ✨ Brand and positioning launch — ClawMaster is now an OpenClaw companion for real life, not just a control plane. New wordmark, Apache 2.0 license, Pride Versioning.

Development

npm install
npm run dev:web       # frontend + backend
npm run dev           # frontend only (port 16223)
npm run dev:backend   # backend only (port 16224)
npm run tauri:dev     # desktop app

Testing and CI

npm test              # unit tests (Vitest)
npm run build         # type check + production build
npm run test:desktop  # desktop smoke (macOS: real Tauri build; Linux/Win: WebDriver)

[!TIP] Run npm test && npm run build before opening a PR — the same steps run in CI.

CI covers core checks including TypeScript, unit tests, and desktop/web build validation.

• Test Suite
• Desktop Bundles

Project layout

clawmaster/
├── packages/web/          React + Vite frontend
├── packages/backend/      Express backend for web mode
├── src-tauri/             Tauri desktop host
├── tests/ui/              YAML-based manual UI flow specs
└── bin/clawmaster.mjs     CLI entry point

Runtime model: Desktop uses Tauri commands; Web mode talks to an Express backend over /api.

Contributing

We warmly welcome more contributions from builders, designers, technical writers, testers, and OpenClaw power users.

If you want to help ClawMaster become more useful for everyday users, please jump in — bug fixes, UX polish, docs improvements, onboarding flows, and future Master Class ideas are all valuable.

Start here:

• AGENTS.md — agent-friendly contributor rules
• CONTRIBUTING.md — setup, testing, commit, and PR guidance
• Ask DeepWiki — explore the repo before changing code

Important

Run npm test locally before opening a PR. Please do not commit generated files or test logs. Node.js is the only permitted runtime — no new language dependencies.

Community: GitHub Discussions · Discord · Feishu

Contributors

Acknowledgments

Project	Role
OpenClaw	Core runtime and configuration model
ClawProbe	Observability daemon
PowerMem · TS SDK	Memory backend
OceanBase seekdb	Retrieval and search workflows
Tauri	Desktop app framework
React	Frontend UI
Vite	Frontend toolchain
Playwright	Browser automation and smoke testing