[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-79744":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":25,"hasPages":23,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":27,"readmeContent":28,"aiSummary":29,"trendingCount":16,"starSnapshotCount":16,"syncStatus":30,"lastSyncTime":31,"discoverSource":32},79744,"PilotDeck","OpenBMB\u002FPilotDeck","OpenBMB","Task-oriented AI Agent productivity platform","https:\u002F\u002Fpilotdeck.openbmb.cn",null,"TypeScript",3015,314,11,16,0,54,1118,1617,292,109.49,"GNU Affero General Public License v3.0",false,"main",true,[],"2026-06-06 04:06:46","\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fbanner.png\" alt=\"PilotDeck\" width=\"680\"\u002F>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  Task-oriented AI Agent productivity platform — redefining operational boundaries and memory evolution, one WorkSpace at a time.\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fpilotdeck.openbmb.cn\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWebsite-pilotdeck.openbmb.cn-FF6B35?style=flat-square&logo=googlechrome&logoColor=white\" alt=\"Official Website\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fpilotdeck.openbmb.cn\u002Fpilotdeck.github.io\u002Fdemo\u002Fp\u002Fpilotdeck-demo\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDemo-Live-brightgreen?style=flat-square\" alt=\"Live Demo\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"LICENSE\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-AGPL_3.0-blue.svg?style=flat-square\" alt=\"License\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fmodelcontextprotocol.io\u002F\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FMCP-Native-6366F1?style=flat-square\" alt=\"MCP Native\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"#-desktop-app\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDesktop-Download-000000?style=flat-square&logo=apple&logoColor=white\" alt=\"Desktop App\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck\u002Fstargazers\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FOpenBMB\u002FPilotDeck?style=flat-square\" alt=\"Stars\"\u002F>\u003C\u002Fa>\n  \u003Cbr\u002F>\n  \u003Ca href=\"#-community\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-Join_Community-5865F2?style=for-the-badge&logo=discord&logoColor=white\" alt=\"Discord\"\u002F>\u003C\u002Fa>\n   \n  \u003Ca href=\"#-community\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FFeishu-Community-00D6B9?style=for-the-badge&logo=bytedance&logoColor=white\" alt=\"Feishu\"\u002F>\u003C\u002Fa>\n   \n  \u003Ca href=\"#-community\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWeChat-Community-07C160?style=for-the-badge&logo=wechat&logoColor=white\" alt=\"WeChat\"\u002F>\u003C\u002Fa>\n  \u003Cbr\u002F>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Cb>English\u003C\u002Fb> | \u003Ca href=\".\u002FREADME.zh.md\">简体中文\u003C\u002Fa>\n  \u003Cbr\u002F>\n  \u003Ca href=\"https:\u002F\u002Fpilotdeck.openbmb.cn\">Website\u003C\u002Fa> · \u003Ca href=\"https:\u002F\u002Fpilotdeck.openbmb.cn\u002Fpilotdeck.github.io\u002Fdemo\u002Fp\u002Fpilotdeck-demo\">Live Demo\u003C\u002Fa> · \u003Ca href=\"https:\u002F\u002Fpilotdeck.openbmb.cn\u002Fpilotdeck.github.io\u002Fdocs\u002Fen\u002Fintroduction\">Tutorial\u003C\u002Fa> · \u003Ca href=\"#-installation--quick-start\">Quick Start\u003C\u002Fa> · \u003Ca href=\"#-key-highlights\">Highlights\u003C\u002Fa> · \u003Ca href=\"#use-cases\">Use Cases\u003C\u002Fa> · \u003Ca href=\"#-community\">Community\u003C\u002Fa>\n\u003C\u002Fp>\n\n---\n\n**News** 🔥\n\n- **[2026.05.28]** PilotDeck is now open source! Visit our official website at [pilotdeck.openbmb.cn](https:\u002F\u002Fpilotdeck.openbmb.cn). We welcome contributions, feedback, and stars from the community.\n\n---\n\n## 💡 About PilotDeck\n\n**PilotDeck** is an open-source agent operating system designed around the concept of \"WorkSpace\". It is jointly developed and open-sourced by Tsinghua University [THUNLP](https:\u002F\u002Fnlp.csai.tsinghua.edu.cn\u002F), [ModelBest](https:\u002F\u002Fmodelbest.cn\u002F), [OpenBMB](https:\u002F\u002Fwww.openbmb.cn\u002F), and [AI9Stars](https:\u002F\u002Fgithub.com\u002FAI9Stars). Targeting general-purpose, multi-task scenarios, PilotDeck is built to be a true *productivity tool* for the Agent era.\n\nA wave of excellent AI Agent harnesses has emerged in recent years, each with its own focus: **Claude Code \u002F Cursor \u002F Trae Solo** brought model reasoning deep into the programming IDE; **Claude Cowork** introduced the notion of project-level isolation to desktop-side knowledge work; **WorkBuddy** connected agents to IM ecosystems such as WeCom and Feishu so AI is one message away.\n\nWhen we shift the lens from \"one-shot programming\" or \"immediate Q&A\" to **long-running, multi-project productivity work**, however, several questions remain open:\n\n- When many projects run in parallel, can memory be **white-box and traceable**? When the AI gets something wrong, can you pinpoint which memory entry caused it and edit it directly — without starting a new chat from scratch?\n- Can token cost be **tracked per task**, so that running agents in the background actually becomes economically viable?\n- Can tasks of different difficulty **automatically be matched to different models**, instead of burning the flagship model on trivial calls?\n- When you step away from the keyboard, can the work keep moving? Can the agent **proactively discover what's worth doing, report progress, and land results as files on disk**?\n\nPilotDeck is an incremental exploration around exactly these questions. It uses the WorkSpace as the fundamental unit — completely isolating files, memory and skills per project — and pairs it with three pillar capabilities: **White-box Memory**, **Smart Routing** and **Always-on**. The entire system natively supports the [Model Context Protocol (MCP)](https:\u002F\u002Fmodelcontextprotocol.io\u002F) and behaves consistently across front-ends (Web \u002F CLI \u002F IM).\n\n### ✨ Key Highlights\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Ctd width=\"50%\" valign=\"top\">\n\n**WorkSpace-Level Isolation & Accretion**\n\nEvery project gets its own file system, memory store and skill set. Parallel work no longer interferes with itself, retrieval has a bounded scope, and skills accrete naturally as each task grows — no more global context pollution.\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fworkspace_en.gif\" width=\"100%\" alt=\"WorkSpace isolation demo\"\u002F>\n\u003C\u002Fp>\n\n\u003C\u002Ftd>\n\u003Ctd width=\"50%\" valign=\"top\">\n\n**Traceable White-box Memory**\n\nMemory generation, extraction, storage and retrieval are visible end-to-end. When the AI mis-remembers, you can pinpoint and fix the offending entry. Built-in **Dream Mode** consolidates memory in idle windows, and supports one-click rollback.\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fmemory_en.gif\" width=\"100%\" alt=\"White-box memory demo\"\u002F>\n\u003C\u002Fp>\n\n\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd width=\"50%\" valign=\"top\">\n\n**Smart Routing & Cost Optimization**\n\nTask difficulty is auto-detected; complex calls go to flagship models (e.g. Claude 3.5 Sonnet \u002F GPT-4o), simple ones drop to lighter models. Through on-device \u002F cloud co-orchestration and precise matching, token spend shrinks dramatically without sacrificing quality.\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Frouter.gif\" width=\"100%\" alt=\"Smart routing demo\"\u002F>\n\u003C\u002Fp>\n\n\u003C\u002Ftd>\n\u003Ctd width=\"50%\" valign=\"top\">\n\n**Always-on Background Execution**\n\nPilotDeck breaks the \"you ask, it answers\" loop: after you sign off, the agent keeps discovering candidate tasks, running long-horizon monitors, and finally lands deliverables as local files with a summary report waiting for you.\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fawo_en.gif\" width=\"100%\" alt=\"Always-on execution demo\"\u002F>\n\u003C\u002Fp>\n\n\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n### 📊 Real-world Numbers\n\nThe three pillar capabilities have shown clear advantages in production-grade workflows:\n\n#### 1. Smart Routing — ~70% cost savings on social-media workloads\n\nIn Xiaohongshu-style social-media operations, enabling Smart Routing automatically demotes simple polishing \u002F layout tasks to a sub-agent (e.g. Sonnet 4.5) and only invokes Opus 4.5 at planning checkpoints:\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Cth width=\"22%\" align=\"left\">Setup\u003C\u002Fth>\n\u003Cth width=\"48%\" align=\"left\">Model configuration\u003C\u002Fth>\n\u003Cth width=\"15%\" align=\"left\">Cost\u003C\u002Fth>\n\u003Cth width=\"15%\" align=\"left\">Multiplier\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cb>Smart Routing ON\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd>Opus 4.5 (main) + Sonnet 4.5 (sub)\u003C\u002Ftd>\n\u003Ctd>\u003Cb>$2.83\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd>\u003Cb>1.1×\u003C\u002Fb>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Smart Routing OFF\u003C\u002Ftd>\n\u003Ctd>All Opus 4.5 (main + sub)\u003C\u002Ftd>\n\u003Ctd>$12.58\u003C\u002Ftd>\n\u003Ctd>5.0×\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Monolithic\u003C\u002Ftd>\n\u003Ctd>Single Opus 4.5 long-react (estimated)\u003C\u002Ftd>\n\u003Ctd>$12.20\u003C\u002Ftd>\n\u003Ctd>4.8×\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd colspan=\"4\">\u003Cimg width=\"840\" height=\"1\" alt=\"\"\u002F>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n#### 2. Smart Routing — 1\u002F6 the cost while beating frontier models on hard tasks\n\nThe research team benchmarked 7 complex tasks (multilingual podcast push, multi-source data reports, domain-specific literature review, codebase architecture docs, etc.). The \"strong main + light sub\" routing setup matches or beats the frontier single-model setup at a fraction of the cost:\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Cth width=\"70%\" align=\"left\">Setting\u003C\u002Fth>\n\u003Cth width=\"15%\" align=\"left\">Score\u003C\u002Fth>\n\u003Cth width=\"15%\" align=\"left\">Cost\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>MiniMax-M2.7 single-agent\u003C\u002Ftd>\n\u003Ctd>37.1\u003C\u002Ftd>\n\u003Ctd>$1.90\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Claude Sonnet 4.6 single-agent\u003C\u002Ftd>\n\u003Ctd>69.1\u003C\u002Ftd>\n\u003Ctd>$18.36\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cb>Sonnet 4.6 (main) + MiniMax-M2.7 (sub)\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd>\u003Cb>70.6\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd>\u003Cb>$3.15\u003C\u002Fb>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd colspan=\"3\">\u003Cimg width=\"840\" height=\"1\" alt=\"\"\u002F>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n#### 3. White-box Memory — layout & tone never bleed across projects\n\nIn black-box agents, mixing tasks in a shared context pool inevitably pollutes memory. PilotDeck's WorkSpace-scoped white-box memory addresses this end-to-end:\n\n\u003Ctable width=\"100%\">\n\u003Cthead>\n\u003Ctr>\n  \u003Cth width=\"14%\" align=\"left\">Dimension\u003C\u002Fth>\n  \u003Cth width=\"41%\" align=\"left\">Current AI Agents (black-box)\u003C\u002Fth>\n  \u003Cth width=\"45%\" align=\"left\">PilotDeck (white-box)\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\n\u003Ctr>\n  \u003Ctd>\u003Cb>Visibility\u003C\u002Fb>\u003C\u002Ftd>\n  \u003Ctd>You can't see what the AI remembers, only what it outputs\u003C\u002Ftd>\n  \u003Ctd>View every memory entry: what was stored, when, and which WorkSpace\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n  \u003Ctd>\u003Cb>Control\u003C\u002Fb>\u003C\u002Ftd>\n  \u003Ctd>Once written, memory can't be edited or removed\u003C\u002Ftd>\n  \u003Ctd>Edit \u002F delete entries, pin critical decisions so they don't drift\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n  \u003Ctd>\u003Cb>Traceability\u003C\u002Fb>\u003C\u002Ftd>\n  \u003Ctd>When it goes wrong, you can't find the root cause\u003C\u002Ftd>\n  \u003Ctd>Generation → extraction → storage → retrieval, all auditable\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n  \u003Ctd>\u003Cb>Isolation\u003C\u002Fb>\u003C\u002Ftd>\n  \u003Ctd>One shared pool — projects bleed into each other\u003C\u002Ftd>\n  \u003Ctd>Scoped per WorkSpace; A's memory never reaches B\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n  \u003Ctd>\u003Cb>Reversible\u003C\u002Fb>\u003C\u002Ftd>\n  \u003Ctd>After compression, the original is gone\u003C\u002Ftd>\n  \u003Ctd>Dream-mode supports \u003Cb>one-click rollback\u003C\u002Fb> to the prior state\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\n\u003C\u002Ftable>\n\n---\n\n## 🖥️ UI & Demo\n\nPilotDeck ships an out-of-the-box Web UI with full WorkSpace management, white-box memory editing, and visualization of multi-agent collaboration.\n\n### Use Cases\n\n> All demos below are generated entirely by edge-side models via PilotDeck's Smart Routing — no cloud-side frontier model required.\n\n#### Work Document Generation\n\n> *\"Survey the Chinese LLM application market and turn it into a formal HTML white paper.\"*\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Process\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Result\u003C\u002Fb>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cimg src=\"assets\u002Fen\u002Fppt_en.gif\" width=\"100%\"\u002F>\u003C\u002Ftd>\n\u003Ctd>\u003Cimg src=\"assets\u002Fresult\u002Fppt_result_en.gif\" width=\"100%\"\u002F>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n#### Mini-Game Development\n\n> *\"Walk me through building an iOS AR mini-game Ball Finder in Vibe Coding mode.\"*\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Process\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Result\u003C\u002Fb>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cimg src=\"assets\u002Fen\u002Fiosgame_en.gif\" width=\"100%\"\u002F>\u003C\u002Ftd>\n\u003Ctd align=\"center\">\u003Cimg src=\"assets\u002Fresult\u002Fios_game_result.gif\" width=\"60%\"\u002F>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n#### AI Engineering Platform Development\n\n> *\"Build a low-code embedding fine-tuning platform from scratch.\"*\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Process\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Result\u003C\u002Fb>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cimg src=\"assets\u002Fen\u002Fmodeltraining_en.gif\" width=\"100%\"\u002F>\u003C\u002Ftd>\n\u003Ctd>\u003Cimg src=\"assets\u002Fresult\u002Fmodeltrainingresult_en.gif\" width=\"100%\"\u002F>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n#### Audio-Video Editing & Social Media Operations\n\n> *\"Push this English podcast to a global audience in Chinese \u002F Japanese \u002F French \u002F Korean \u002F Spanish \u002F Arabic.\"*\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Process\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd width=\"50%\" align=\"center\">\u003Cb>Result (with audio)\u003C\u002Fb>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cimg src=\"assets\u002Fen\u002Fpodcast_en.gif\" width=\"100%\"\u002F>\u003C\u002Ftd>\n\u003Ctd>\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fa7245467-ee3c-4939-a055-c56576ac56d1\n\n\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n---\n\n## 📦 Installation & Quick Start\n\nWe provide a one-line installer for macOS \u002F Linux, plus a source-based workflow for developers.\n\n### Option A: One-line install (recommended, macOS \u002F Linux)\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002FOpenBMB\u002FPilotDeck\u002Fmain\u002Finstall.sh | bash\n```\n\nThe script auto-installs Node.js 22, clones the repo, installs dependencies, and builds the frontend. Once it finishes:\n\n```bash\npilotdeck            # starts the server at http:\u002F\u002Flocalhost:3001\npilotdeck status     # check runtime status\n```\n\n### Option B: From source (for developers)\n\n**1. Clone and install dependencies**\n\n> This repo uses [Git LFS](https:\u002F\u002Fgit-lfs.com\u002F) for large media assets. Make sure `git lfs` is installed before cloning.\n> If you don't need the demo videos\u002FGIFs, add `GIT_LFS_SKIP_SMUDGE=1` before `git clone` to skip downloading them.\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck.git\ncd PilotDeck\n\nnpm install              # root deps (Gateway runtime)\ncd ui && npm install     # UI deps\ncd ..\n```\n\n**2. Configure a model provider**\n\nPilotDeck reads `~\u002F.pilotdeck\u002Fpilotdeck.yaml`. You can create it manually, let the bootstrap script generate one, **or just open the Web UI and configure providers visually in the settings panel.**\nSupported protocols include OpenAI, Anthropic, DeepSeek, Qwen, Kimi, MiniMax and other OpenAI-compatible endpoints.\n\n```yaml\nschemaVersion: 1\nagent:\n  model: deepseek\u002Fdeepseek-v4-pro\nmodel:\n  providers:\n    deepseek:\n      protocol: openai\n      url: https:\u002F\u002Fapi.deepseek.com\u002Fv1\n      apiKey: sk-your-api-key\n```\n\n**3. Start the services**\n\n```bash\ncd ui && npm run dev     # dev mode (HMR), visit http:\u002F\u002Flocalhost:5173\n# or\ncd ui && npm run start   # production mode, visit http:\u002F\u002Flocalhost:3001\n```\n\n### Option C: Docker Compose\n\nIf Docker is installed, you can start PilotDeck with:\n\n```bash\ndocker compose up -d\n```\n\n### 🖥️ Desktop App\n\nPre-built installers — double-click to run, no command-line setup required.\n\n| Platform | Download |\n|----------|----------|\n| macOS (Apple Silicon) | [PilotDeck-latest-arm64.dmg](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck\u002Freleases\u002Flatest\u002Fdownload\u002FPilotDeck-latest-arm64.dmg) |\n| Windows (x64) | [PilotDeck-latest-win-x64.exe](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck\u002Freleases\u002Flatest\u002Fdownload\u002FPilotDeck-latest-win-x64.exe) |\n| Windows (ARM64) | [PilotDeck-latest-win-arm64.exe](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck\u002Freleases\u002Flatest\u002Fdownload\u002FPilotDeck-latest-win-arm64.exe) |\n\n> macOS DMG is signed and Apple-notarized. If you received the DMG via Feishu\u002FWeChat\u002FQQ, see [INSTALL.md](apps\u002Fdesktop\u002FINSTALL.md) for a one-command fix.\n>\n> Browse all versions: [GitHub Releases](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck\u002Freleases\u002Flatest)\n\n---\n\n## 🛠️ Extension Protocol\n\nPilotDeck has an open plugin architecture with a strict boundary between the open-source core and plugin customization. Extending the system is a `plugin.json` away:\n\n- **MCP Servers** — first-class integration with any Model Context Protocol server.\n- **Tools & Skills** — register custom tools, or pull community skills via [ClawHub](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002Fclawhub).\n- **Lifecycle Hooks** — intercept `PreToolUse`, `UserPromptSubmit`, and other critical lifecycle events.\n- **Custom Memory** — plug in your own memory store provider.\n\n---\n\n## 🤝 Contributing\n\nThanks to everyone who has contributed code, feedback, and ideas. New contributors are warmly welcome — let's build the next-gen agent OS together.\n\nWorkflow: **Fork → feature branch → PR**.\n\n---\n\n## 💬 Community\n\n- For bugs and feature requests, please open a [GitHub Issue](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck\u002Fissues).\n- Join our community channels:\n\n\u003Ctable width=\"100%\">\n\u003Ctr>\n\u003Ctd width=\"33%\" align=\"center\">\u003Cb>WeChat Community\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd width=\"33%\" align=\"center\">\u003Cb>Feishu Community\u003C\u002Fb>\u003C\u002Ftd>\n\u003Ctd width=\"33%\" align=\"center\">\u003Cb>Discord Community\u003C\u002Fb>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd align=\"center\">\u003Cimg src=\"assets\u002Fcommunity\u002Fqr-wechat.png\" width=\"200\" alt=\"WeChat QR\"\u002F>\u003C\u002Ftd>\n\u003Ctd align=\"center\">\u003Cimg src=\"assets\u002Fcommunity\u002Fqr-feishu.png\" width=\"200\" alt=\"Feishu QR\"\u002F>\u003C\u002Ftd>\n\u003Ctd align=\"center\">\u003Cimg src=\"assets\u002Fcommunity\u002Fqr-discord.png\" width=\"200\" alt=\"Discord QR\"\u002F>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n---\n\n## 🙏 Acknowledgements\n\nWe thank Agent OS pioneers such as OpenClaw, Claude Code, Codex, Cursor, and Hermes for their explorations that helped shape this field.\n\nPilotDeck builds upon the following outstanding open-source projects:\n\n- [ClawXRouter](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FClawXRouter) — Intelligent model routing\n- [ClawXMemory](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FClawXMemory) — Agent memory system\n- [Claude Code UI](https:\u002F\u002Fgithub.com\u002Fsiteboon\u002Fclaudecodeui) — Web UI reference\n- [Claude Code Router](https:\u002F\u002Fgithub.com\u002Fmusistudio\u002Fclaude-code-router) — Model routing reference\n- [UltraRAG](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FUltraRAG) — RAG framework\n- [Anthropic Skills](https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fskills) — Agent skill framework and built-in skills (skill-creator)\n- [Vercel Labs Skills](https:\u002F\u002Fgithub.com\u002Fvercel-labs\u002Fskills) — find-skills skill\n- [MiniMax-AI Skills](https:\u002F\u002Fgithub.com\u002FMiniMax-AI\u002Fskills) — minimax-pdf skill\n- [frontend-slides](https:\u002F\u002Fgithub.com\u002Fzarazhangrui\u002Ffrontend-slides) — Create beautiful slides on the web using a coding agent's frontend skills\n- [Karpathy Guidelines](https:\u002F\u002Fx.com\u002Fkarpathy\u002Fstatus\u002F2015883857489522876) — LLM coding behavioral guidelines\n- [Vite](https:\u002F\u002Fgithub.com\u002Fvitejs\u002Fvite) — Frontend build tool\n- [React](https:\u002F\u002Fgithub.com\u002Ffacebook\u002Freact) — UI framework\n- [Tailwind CSS](https:\u002F\u002Fgithub.com\u002Ftailwindlabs\u002Ftailwindcss) — Utility-first CSS framework\n- [shadcn\u002Fui](https:\u002F\u002Fgithub.com\u002Fshadcn-ui\u002Fui) — Accessible component primitives for React\n\n---\n\n## 🏢 Joint Development\n\nPilotDeck is jointly developed by Tsinghua University [THUNLP](https:\u002F\u002Fnlp.csai.tsinghua.edu.cn\u002F), [ModelBest](https:\u002F\u002Fmodelbest.cn\u002F), [OpenBMB](https:\u002F\u002Fwww.openbmb.cn\u002F) and [AI9Stars](https:\u002F\u002Fgithub.com\u002FAI9Stars).\n\n---\n\n## ⭐ Support Us\n\nIf PilotDeck has been helpful in your work or research, please consider giving us a Star on GitHub!\n\n---\n\n## 📝 Citation\n\n```bibtex\n@misc{pilotdeck2026,\n  author       = {PilotDeck Team},\n  title        = {PilotDeck: A WorkSpace-Centric Open-Source Agent Operating System},\n  howpublished = {\\url{https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FPilotDeck}},\n  year         = {2026},\n  note         = {Accessed: 2026-05-29}\n}\n```\n\n## 📄 License\n\nThis project is licensed under the [GNU Affero General Public License v3.0](LICENSE).\n","PilotDeck 是一个面向任务的AI代理生产力平台，旨在通过“WorkSpace”概念重新定义操作边界和记忆进化。该项目采用TypeScript开发，支持多任务场景下的高效协作与管理，并且原生支持Model Context Protocol (MCP)以增强模型间的交互能力。它适合需要集成多种AI工具、提高工作效率的企业或个人开发者使用。此外，PilotDeck 提供了桌面应用程序下载以及在线演示版本，方便用户快速上手体验其功能。",2,"2026-06-06 03:59:20","CREATED_QUERY"]