🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman
Desktop pet monitor for AI coding agents
Vibe Pet connects to agent hooks and session data from tools like Codex, Cursor, Windsurf, and Claude Code. It shows each agent’s state as a pet card on the desktop and can send the same status to supported hardware such as Wio Terminal or ESP32-S3 devices.

Builders who want their agent activity shown as a live desktop and hardware companion.
You can see what your agents are doing at a glance instead of checking logs or switching windows.
What it does
Agent state monitoring
Tracks live activity from supported agents and maps states like thinking, tool use, approval waiting, completed, and error.
Multi-agent pet view
Shows each active editor or agent as its own pet card instead of combining everything into one status.
Desktop pets
Spawns pets on the desktop, where each one can be dragged independently.
Hardware sync
Sends state over BLE to devices such as Wio Terminal, SenseCAP Indicator, and ESP32-S3-based boards.
Character switching
Lets you pick from Petdex characters or use a custom character you made.
Broad agent integrations
Supports hooks, plugin-style integrations, and session monitoring for many agent tools, including Codex, Cursor, Windsurf, Claude Code, Gemini CLI, and Copilot CLI.
How to get it
- 1This is a Gatekeeper block (unsigned or not notarized build), not a runtime crash.…
xattr -dr com.apple.quarantine "/Applications/Vibe Pet.app"
- 2If needed, re-sign locally
sudo codesign --force --deep --sign - "/Applications/Vibe Pet.app"
README
Vibe Pet
Hardware desktop pets for your AI coding agents
Choose from thousands of characters or create your own in one click
Turn Codex, Cursor, Windsurf, and other agents into live companions across desktop and hardware
English · Protocol · Petdex · 中文 · 日本語
Vibe Pet is a hardware desktop pet project for AI coding agents. It watches live activity from Codex, Cursor, Windsurf, and other AI coding agents in CLIs and IDEs, turns states such as thinking, tool use, waiting for approval, completed, and error into animated pets, then syncs the same state to Wio Terminal or ESP32-S3 hardware over BLE.
Highlights
- Multi-agent pet view: every active editor or AI coding agent gets its own pet card instead of sharing one combined state.
- Real desktop pets: pets from the main window are also spawned onto the desktop, and each one can be dragged independently.
- Hardware sync: send state over Bluetooth to devices such as Wio Terminal or ESP-AI-MINI AI Dev Kit.
- Character switching: choose a character from Petdex by default or use a custom character you made.
Supported Agents
Vibe Pet automatically tries to install or sync integrations for:
| Agent | Integration |
|---|---|
| Codex | Hooks and JSONL session monitoring |
| Cursor | Hooks |
| Windsurf | Cascade hooks |
| Claude CLI | Hooks |
| Claude Code | Hooks |
| Gemini CLI | Hooks |
| Copilot CLI | Hooks |
| CodeBuddy | Hooks |
| Kimi Code CLI | Hooks |
| Qwen Code | Hooks |
| OpenClaw | Plugin |
| opencode | Plugin |
| Qoder | Hooks |
| Hermes Agent | Plugin |
| Reasonix CLI | Hooks |
Supported Hardware
| Hardware | Display / role | Sync | Adapted | Notes |
|---|---|---|---|---|
| SenseCAP Indicator | 4-inch 480×480 touch display | BLE | ✅️ | ESP32-S3 + RP2040; ST7701S RGB display; LVGL character rendering. |
| Wio Terminal | Main animated pet display | BLE | Ing | Mature BLE animated display target. |
| Vibe Pet | ESP32-S3-based AI dev kit | BLE | ✅️ | TFT target; LVGL character rendering. |
Want to bring Vibe Pet to your own device? The BLE and Wi-Fi hardware payloads are intentionally small, so new screens, status lights, badges, and custom boards can be added without changing the desktop app.
Download and Install
Download the installer for your platform from the Releases page.
- macOS: download the
.dmgor.zipbuild. - Windows: download the
.exeinstaller. - Linux: download the
.AppImageor.debpackage.
After installation, launch Vibe Pet and use the desktop app to connect your device. You can also use Vibe Pet without hardware if you only want the desktop pets.
FAQ
macOS says "Vibe Pet" is damaged and can't be opened
This is a Gatekeeper block (unsigned or not notarized build), not a runtime crash. Remove the quarantine attribute and try again:
xattr -dr com.apple.quarantine "/Applications/Vibe Pet.app"
If needed, re-sign locally:
sudo codesign --force --deep --sign - "/Applications/Vibe Pet.app"
Technical Documentation
- Project structure, local endpoints, BLE / Wi-Fi behavior, and hook mapping are documented in AGENT.MD
- Protocol documentation starts at the protocol index
- English protocol docs:
- Chinese protocol docs:
Sponsor
Contributing
Ideas, bug reports, hardware ports, new agent integrations, translations, and UI improvements are all welcome. If you want to contribute code, read CONTRIBUTING.md for development setup, packaging, testing, and PR guidance.
Files in the repo
- .github
- docs
- images
- scripts
- src
- .gitignore
- AGENT.MD
- CONTRIBUTING.md
- electron-builder.json
- LICENSE
- package-lock.json
- package.json
- README.ja-JP.md
- README.md
- README.zh-CN.md
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More tools
The best-benchmarked open-source AI memory system. And it's free.
Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io
Never stop coding. Free MIT AI gateway: one endpoint, 352 providers (150+ free), 1200+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 550+ contributors
Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.