🌐 Language: 中文 | English | 日本語

A pixel-art AI office dashboard — visualize your AI assistant's work status in real time, so you can see at a glance who's doing what, what they did yesterday, and whether they're online.

Supports multi-agent collaboration, trilingual UI (CN/EN/JP), AI-powered room design, and desktop pet mode. Best experienced with OpenClaw, but also works standalone as a status dashboard.

This project was co-created by Ring Hyacinth and Simon Lee, and is continuously maintained and improved together with community contributors (@Zhaohan-Wang, @Jah-yee, @liaoandi). Issues and PRs are welcome — thank you to everyone who contributes.

---

If you're using OpenClaw, just send this to your lobster:

Please follow this SKILL.md to deploy Star Office UI for me:
https://github.com/ringhyacinth/Star-Office-UI/blob/master/SKILL.md

Your lobster will automatically clone the repo, install dependencies, start the backend, configure status sync, and send you the access URL.

Requires Python 3.10+ (the codebase uses X | Y union type syntax, which is not supported on 3.9 or earlier)

# 1) Clone the repo
git clone https://github.com/ringhyacinth/Star-Office-UI.git
cd Star-Office-UI

# 2) Install dependencies (Python 3.10+ required)
python3 -m pip install -r backend/requirements.txt

# 3) Initialize state file (first run)
cp state.sample.json state.json

# 4) Start the backend
cd backend
python3 app.py

Open http://127.0.0.1:19000 and try switching states:

python3 set_state.py writing "Organizing documents"
python3 set_state.py error "Found an issue, debugging"
python3 set_state.py idle "Standing by"

---

This is the full experience. Your agent automatically switches status as it works, and the pixel character walks to the corresponding office area in real time — just open the page and see what your AI is doing right now.

You can still deploy and use it. You can:

• Use set_state.py or the API to push status manually or via scripts
• Use it as a pixel-art personal status page or remote work dashboard
• Connect any system that can send HTTP requests to drive the status

---

1. Status Visualization — 6 states (idle / writing / researching / executing / syncing / error) mapped to different office areas with animated sprites and speech bubbles
2. Yesterday Memo — Automatically reads the latest daily log from memory/*.md, sanitizes it, and displays it as a "Yesterday Memo" card
3. Multi-Agent Collaboration — Invite other agents to join your office via join keys and see everyone's status in real time
4. Trilingual UI — Switch between Chinese, English, and Japanese with one click; all UI text, bubbles, and loading messages update instantly
5. Custom Art Assets — Manage characters, scenes, and decorations through the sidebar; dynamic frame sync prevents flickering
6. AI-Powered Room Design — Connect your own Gemini API to generate new office backgrounds; core features work fine without an API
7. Mobile-Friendly — Open on your phone for a quick status check on the go
8. Security Hardening — Sidebar password protection, weak-password blocking in production, hardened session cookies
9. Flexible Public Access — Use Cloudflare Tunnel for instant public access, or bring your own domain / reverse proxy
10. Desktop Pet Mode — Optional Electron desktop wrapper that turns the office into a transparent desktop widget (see below)

---

cd Star-Office-UI
python3 -m pip install -r backend/requirements.txt

cp state.sample.json state.json

cd backend
python3 app.py

Open http://127.0.0.1:19000

✅ For local development you can start with the defaults; in production, copy .env.example to .env and set strong random values for FLASK_SECRET_KEY and ASSET_DRAWER_PASS to avoid weak passwords and session leaks.

python3 set_state.py writing "Organizing documents"
python3 set_state.py syncing "Syncing progress"
python3 set_state.py error "Found an issue, debugging"
python3 set_state.py idle "Standing by"

cloudflared tunnel --url http://127.0.0.1:19000

Share the https://xxx.trycloudflare.com link with anyone.

python3 scripts/smoke_test.py --base-url http://127.0.0.1:19000

If all checks report OK, your deployment is good to go.

---

The following section is for OpenClaw users. If you don't use OpenClaw, feel free to skip this.

Add the following rule to your SOUL.md (or agent config) so your agent updates its status automatically:

## Star Office Status Sync Rules
- When starting a task: run `python3 set_state.py <state> "<description>"` before beginning work
- When finishing a task: run `python3 set_state.py idle "Standing by"` before replying

6 states → 3 office areas:

State	Office Area	When to use
idle	🛋 Breakroom (sofa)	Standing by / task complete
writing	💻 Workspace (desk)	Writing code or docs
researching	💻 Workspace	Searching / researching
executing	💻 Workspace	Running commands / tasks
syncing	💻 Workspace	Syncing data / pushing
error	🐛 Bug Corner	Error / debugging

Step 1: Prepare join keys

When you start the backend for the first time, if there is no join-keys.json in the project root, the service will automatically create one based on join-keys.sample.json (which contains an example key such as ocj_example_team_01). You can then edit the generated join-keys.json to add, modify, or remove keys; by default each key supports up to 3 concurrent users.

Step 2: Have the guest run the push script

The guest only needs to download office-agent-push.py and fill in 3 variables:

JOIN_KEY = "ocj_starteam02"          # The key you assign
AGENT_NAME = "Alice's Lobster"       # Display name
OFFICE_URL = "https://office.hyacinth.im"  # Your office URL

python3 office-agent-push.py

The script auto-joins and pushes status every 15 seconds. The guest will appear on the dashboard, moving to the appropriate area based on their state.

Step 3 (optional): Guest installs a Skill

Guests can also use frontend/join-office-skill.md as a Skill — their agent will handle setup and pushing automatically.

See frontend/join-office-skill.md for full guest onboarding instructions.

---

Endpoint	Description
GET /health	Health check
GET /status	Get main agent status
POST /set_state	Set main agent status
GET /agents	List all agents
POST /join-agent	Guest joins the office
POST /agent-push	Guest pushes status
POST /leave-agent	Guest leaves
GET /yesterday-memo	Get yesterday's memo
GET /config/gemini	Get Gemini API config
POST /config/gemini	Set Gemini API config
GET /assets/generate-rpg-background/poll	Poll image generation progress

---

The desktop-pet/ directory contains a Electron-based desktop wrapper that turns the pixel office into a transparent desktop widget.

cd desktop-pet
npm install
npm run dev

• Auto-launches the Python backend on startup
• Window points to http://127.0.0.1:19000/?desktop=1 by default
• Customizable via environment variables for project path and Python path

⚠️ This is an optional, experimental feature, primarily developed and tested on macOS. See desktop-pet/README.md for details.

🙏 The desktop pet module was independently developed by @Zhaohan-Wang — thank you for this contribution!

---

Guest character animations use free assets by LimeZu:

• Animated Mini Characters 2 (Platformer) [FREE]

Please keep attribution when redistributing or demoing, and follow the original license terms.

• Code / Logic: MIT (see LICENSE)
• Art Assets: Non-commercial use only (learning / demo / sharing)

For commercial use, replace all art assets with your own original artwork.

---

Date	Summary	Details
2026-03-06	🔌 Default port updated — backend default port changed from 18791 to 19000 to avoid conflicts with OpenClaw Browser Control; synced scripts, desktop shells, and docs defaults	docs/CHANGELOG_2026-03.md
2026-03-05	📱 Stability fixes — CDN cache fix, async image generation, mobile sidebar UX, join key expiration & concurrency	docs/UPDATE_REPORT_2026-03-05.md
2026-03-04	🔒 P0/P1 Security hardening — weak password blocking, backend refactor, stale-state auto-idle, skeleton loading	docs/UPDATE_REPORT_2026-03-04_P0_P1.md
2026-03-03	📋 Open-source release checklist completed	docs/OPEN_SOURCE_RELEASE_CHECKLIST.md
2026-03-01	🎉 v2 Rebuild — Trilingual support, asset management system, AI room design, full art asset overhaul	docs/FEATURES_NEW_2026-03-01.md

---

Star-Office-UI/
├── backend/            # Flask backend
│   ├── app.py
│   ├── requirements.txt
│   └── run.sh
├── frontend/           # Frontend pages & assets
│   ├── index.html
│   ├── join.html
│   ├── invite.html
│   └── layout.js
├── desktop-pet/        # Electron desktop wrapper (optional)
├── docs/               # Documentation & screenshots
│   └── screenshots/
├── office-agent-push.py  # Guest push script
├── set_state.py          # Status switch script
├── state.sample.json     # State file template
├── join-keys.sample.json # Join key template (runtime generates join-keys.json)
├── SKILL.md              # OpenClaw Skill
└── LICENSE               # MIT License

---