Show HN: PearlOS：我们为 AI 提供了一个会说话的桌面环境，而不是一个文本框
2 分•作者: stephanieriggs•4 天前
AI 在终端中非常强大，但许多人对使用命令行并不熟悉——这意味着他们无法访问或充分利用 AI 的全部力量。
因此，我们设想了一种不同的体验，并基于以下问题：如果 AI 拥有整个桌面来工作会怎样？如果 AI 成为体验本身而不是工具会怎样？如果这种体验既有趣、令人愉悦又直观会怎样？今天，我们发布了我们创建内容的预览视频。
PearlOS 是一个基于浏览器的桌面环境，AI 助手（Pearl）与你对话，打开应用程序、管理窗口、构建角色、进行研究、做笔记、搜索网络，并控制整个用户界面。你只需和她对话，屏幕上就会发生各种事情。
目前还处于早期阶段。但核心功能已经可用，我们整个（小型）团队都在日常任务中使用它。
演示视频：[https://www.youtube.com/watch?v=aKO52ox0dx0](https://www.youtube.com/watch?v=aKO52ox0dx0)
GitHub：[https://github.com/NiaExperience/PearlOS/](https://github.com/NiaExperience/PearlOS/)
我们希望将其构建成“个人珍珠”，让每个人都能拥有，无需任何编码/AI 经验。
我们正在寻求架构反馈、贡献者和诚实的批评。
打开它后，你将获得：
* 一个带有应用程序的桌面环境（笔记、浏览器、YouTube、文件管理器、计算器、音乐播放器），所有这些都
* 一个你可以自然对话的声音。处理打断、轮流对话、真正的交流。
* 持久记忆。Pearl 会记住你的项目、偏好和跨会话的过去对话。她会接续你之前的工作。
* 子代理群。Pearl 可以将复杂的任务分解成并行代理作业，这些作业在后台运行，而你继续对话。
技术栈：
* Next.js 前端（桌面在浏览器中运行）
* 多模型路由（用于聊天的快速模型，用于复杂推理的更重模型，可配置）
* Pipecat 用于实时语音（Deepgram STT，PocketTTS 用于本地 TTS）
* OpenClaw 用于代理编排
* 使用你自己的 API 密钥，可以替换任何 LLM
它不是什么：
* 不是另一个 ChatGPT 包装器
* 不是一个附加了 AI 的终端
* 不收集你的数据（在你的硬件上运行，没有遥测）
感谢大家！
Stephanie & PearlOS 团队
查看原文
AI is awesome in the terminal but many people aren't comfortable using command lines - which means they can't access or utilize the full power of AI.<p>So we envisioned a different kind of experience, and rooted it in questions like: What if the AI had an entire desktop to work with? What if AI was the experience rather than the tool? And what if that experience was fun, delightful, and intuitive? Today we dropped a first look video of what we created.<p>PearlOS is a browser based desktop environment where the AI companion (Pearl) talks to you to open apps, manage windows, build characters, research, takes notes, searches the web, and controls the whole UI. You just talk to her and things happen on screen.<p>It's early. But the core works and our entire (small) team uses it for daily tasks
Walkthrough Video: <a href="https://www.youtube.com/watch?v=aKO52ox0dx0" rel="nofollow">https://www.youtube.com/watch?v=aKO52ox0dx0</a>
GitHub: <a href="https://github.com/NiaExperience/PearlOS/" rel="nofollow">https://github.com/NiaExperience/PearlOS/</a><p>We want to build this out into “personal pearls” that everyone can have with no coding/AI experience at all needed.<p>Looking for architecture feedback, contributors, and honest criticism.<p>What you get when you open it: * A desktop environment with apps (notes, browser, YouTube, file manager, calculator, music player), all * A voice you can talk to naturally. Interruption handling, turn taking, real conversation. * Persistent memory. Pearl remembers your projects, preferences, and past conversations across sessions. She picks up where you left off. * Sub-agent swarms. Pearl can break complex tasks into parallel agent jobs that run in the background while you keep talking.<p>Stack: * Next.js frontend (the desktop runs in a browser) * Multi-model routing (fast model for chat, heavier model for complex reasoning, configurable) * Pipecat for real-time voice (Deepgram STT, PocketTTS for local TTS) * OpenClaw for agent orchestration * Bring your own API keys, swap in any LLM<p>What it's not: * Not another ChatGPT wrapper * Not a terminal with an AI bolted on * Not collecting your data (runs on your hardware, no telemetry)<p>Thanks everyone!
Stephanie & the PearlOS team