aicodeking/pandacode

PandaCode – A native desktop workspace for AI coding agents

What it is

• An open‑source desktop application (macOS, Windows, Linux) built with Tauri 2, Rust and TypeScript.
• Provides a single window where you can run many coding CLIs (Claude Code, Codex, Kimi Code, OpenCode, GLM Coding Plan, Xiaomi MiMo, Antigravity, xAI Grok Build, Pi, DeepSeek, etc.) alongside your code, Git, a terminal and a web browser.
• Stores all data locally in a SQLite database; no accounts, telemetry, or forced cloud services.

Key features

• Multi‑harness support – pick any installed coding CLI, model and reasoning effort per task.
• Task board – visual Kanban view (Pending → In‑progress → Review → Completed) with queueing, approvals and turn‑by‑turn diff tracking.
• Agent teams – an architect agent can spawn up to three worker agents, mixing different harnesses.
• Bots – expose an agent as a Telegram, Slack, Discord, WhatsApp or web chat; each conversation appears as a normal task.
• Local & cloud sandboxes – drive head‑less macOS/Linux VMs (via Tart) or cloud sandboxes (Daytona, Cloudflare, AWS) from the UI.
• Simulators – run iOS Simulator or Android emulator side‑by‑side with the chat.
• One‑click deploys – trigger Vercel, Cloudflare, Railway, AWS or Google Cloud deploys with approval‑gated CLI runs.
• Deep Git integration – view repos, PRs, issues; stage, commit, create worktrees, merge into main, and launch gh‑based actions.
• Quick launcher – global shortcut (⌘/Ctrl + Shift + Space) to start or switch tasks from anywhere.
• Token saver – per‑task response‑style packs to reduce token usage.
• Attachments & mockup sketcher – drop images, PDFs, folders, or draw UI mockups directly into a task.
• Privacy‑first – only contacts the services you configure; keys are kept in the OS keychain; minimal public‑info checks for updates.

Installation

• Download signed binaries from the GitHub Releases page (macOS arm64 DMG, Windows x64 EXE, Linux x64 AppImage).
• Or run npx pandacode to fetch, verify and install the latest build.
• After install the app updates itself in the background (can be disabled in Settings → About).

Building from source

git clone https://github.com/aicodeking/pandacode.git
cd pandacode
npm ci               # install Node deps
npm run desktop       # start Tauri dev window (or npm run tauri build for a release)

Requires Node 22, Rust stable and the Tauri prerequisites for your OS.

License

• Distributed under the PandaCode License: MIT‑style with an attribution clause – forks used commercially or publicly must state they are a fork of PandaCode and link back to the repo.

Who it’s for

• Developers who want a local, integrated UI to orchestrate multiple AI coding assistants, manage their code with Git, and run bots or cloud sandboxes without leaving the desktop.
• Teams that need visual task tracking and the ability to combine different LLM providers in a single workflow.

---

All details are taken directly from the repository’s README; no additional features are inferred.