Show HN: Yogurt – Free, private, open-source alternative to Granola
282 points
• 1 day ago
• Article
Link
Yogurt 是一款免费的开源应用，旨在生成基于 AI 的会议纪要，同时将用户隐私放在首位。与许多其他转录服务不同，它不会以参与者机器人的身份加入会议。相反，它会在本地运行：录制你的麦克风和系统音频，确保所有转录、摘要以及 AI 处理都直接保存在你自己的电脑上。这样，用户无需创建账号，也不必依赖外部云服务进行处理，就能对会议数据保持完全掌控。
这款应用的工作方式是捕获通话音频，并在对话展开的同时提供实时转录。为实现这一点，Yogurt 使用 Whisper.cpp 进行转录；在受支持的 Mac 上，还可以借助 Apple SpeechAnalyzer 进行实时词语识别。用于摘要和提问时，它会调用本地的开放权重模型（如 Gemma 、 Gpt-oss 、 Qwen3 等系列），并通过 Llama.cpp 执行。初次模型下载完成后，这些强大功能即可完全离线运行，让用户能够搜索会议内容、复制笔记，或将信息导出为 Markdown 文件。
该平台的一项关键能力是支持多语言任务，提供 99 种语言。它会通过监测麦克风使用情况或日历事件来自动判断通话何时开始，但它始终保持低调：只有在用户明确选择要录制时才会开始。 macOS 上，应用还可以在用户授予必要的辅助功能权限后，进一步识别来自 Zoom 、 Teams 、 Meet 等平台的发言者姓名。
为方便与其他工作流集成，Yogurt 提供一个本地的 Model Context Protocol（MCP）服务器。这样，其他助手类应用就可以读取你保存的笔记和转录内容。不过该功能默认处于关闭状态，需要手动启用。启用后，访问权限被限制为只读模式；应用还通过本地套接字或命名管道的权限设置来确保安全，从而将访问严格限定在你的账户范围内。
Yogurt 支持 macOS 、 Windows 和 Linux，想要获得高效运行体验，至少需要 8 GB 内存。由于它完全在设备端运行，并且避免采用"会议中的机器人"模式，因此为需要整理会议文档、又不想牺牲对话隐私的专业人士提供了一种安全且透明的解决方案。所有源代码均以 MIT 许可证发布，进一步体现了该项目在可访问性与社区驱动开发方面的承诺。
Yogurt is a free, open-source application designed to generate AI-powered meeting notes while prioritizing user privacy. Unlike many other transcription services, it does not join meetings as a participant bot. Instead, it operates locally by recording your microphone and system audio, ensuring that all transcriptions, summaries, and AI processing remain directly on your own computer. This approach allows users to maintain full control over their meeting data without needing to create an account or rely on external cloud services for processing.
The application functions by capturing call audio and providing live transcriptions as the conversation unfolds. To achieve this, Yogurt utilizes Whisper.cpp for transcription, while supported Macs can leverage Apple SpeechAnalyzer for live word recognition. For summarization and querying, it uses local, open-weight models from families such as Gemma, Gpt-oss, or Qwen3, which are executed through Llama.cpp. Once the initial model download is complete, these powerful features function entirely offline, allowing users to search through meetings, copy notes, or export information as Markdown files.
One of the platform's key features is its ability to handle multi-language tasks, supporting 99 different languages. While it automatically detects when a call starts by monitoring microphone usage or calendar events, it remains a discreet tool that only records when the user explicitly chooses to do so. On macOS, the app offers additional utility by identifying speaker names from platforms like Zoom, Teams, and Meet, provided the user grants the necessary accessibility permissions.
To integrate with other workflows, Yogurt includes a local Model Context Protocol (MCP) server. This allows other assistant applications to read your saved notes and transcripts, though this feature is disabled by default and requires manual activation. When enabled, access is restricted to read-only mode, and the application ensures security through local socket or named pipe permissions that limit access strictly to the user's account.
Yogurt is compatible with macOS, Windows, and Linux, requiring a minimum of 8 GB of memory for efficient performance. By running entirely on-device and avoiding the "bot-in-the-meeting" model, it offers a secure, transparent solution for professionals who need organized meeting documentation without compromising the privacy of their conversations. All source code is available under the MIT license, reinforcing the project's commitment to accessibility and community-driven development.
108 comments • Comments Link
- Biscotti 提供了原生的 macOS 会议录制能力，功能完全本地化，避免了 Electron 带来的内存开销，同时实现对 Windows 和 Linux 的跨平台支持。
- 围绕 Granola 等基于云的 AI 录制服务的隐私担忧，推动了开源本地替代方案的快速增长，尽管用户仍在"数据完全由自己掌控"与"云 API 的高性能"之间反复权衡。
- 会议录制工具会运用多种技术来处理音频，比如借助辅助功能权限进行发言人归属，或使用 pyannote 等本地模型来做区分。
- 一些参与者对几乎同质化的笔记应用层出不穷感到沮丧，担心在现有项目的核心功能基本一致的情况下，是否还有必要持续大规模重构。
- 命名规则的争论依然激烈：不少开发者即便可能面临潜在的 SEO 和品牌冲突，仍倾向于采用通用且与食物相关的名称；也有开发者主张使用更独特的标识符。
- 法律与道德合规仍是关键摩擦点，尤其是在双向同意的合规要求上。有建议认为：每当通话中新增成员时，应用应主动向所有参与者发出录音相关的通知。
- 用户越来越重视深度定制，包括自定义总结提示词、配置触发自动化脚本（例如把转录内容同步到 Obsidian），以及手动选择特定 AI 模型的能力。
- 转录效果往往取决于本地硬件性能，这促使用户要求提供更完善的模型选择器，并为总结等高耗能任务提供卸载选项。
- 诸如通过 Unix sockets 实现本地进程间通信等实现细节，会受到社区的严格审查，以确保本地转录访问安全且可验证身份。
- 不少人已经开始用这些工具来梳理会议中的混乱、提升后续跟进质量，并减轻在"主动参与"和"手动记笔记"之间切换所带来的认知负担。
这场讨论反映出社区正在更广泛地转向"本地优先"的 AI 工具，这背后主要是隐私焦虑以及对个人数据进行细粒度控制的需求。尽管会议助理市场逐渐趋于饱和，开发者仍通过不同的技术取舍来拉开差异，例如原生 UI 的性能与跨平台兼容性的权衡，或在"仅本地处理"与"API 驱动的模型灵活性"之间做取舍。尽管围绕命名以及静默录音的伦理问题仍存在争议，但各方的共识十分清晰：这些工具已经成为在复杂、信息密集的工作环境中开展高效生产的必备手段。 • Biscotti provides a native macOS meeting recorder that functions entirely locally and avoids the memory overhead of Electron, while offering cross-platform support for Windows and Linux.
• Privacy concerns regarding cloud-based AI recording services like Granola have fueled a surge in open-source, local alternatives, though users remain divided on the trade-offs between full data sovereignty and the superior performance of cloud APIs.
• Meeting recording tools utilize various techniques to handle audio, such as leveraging accessibility permissions for speaker attribution or employing local models like pyannote for diarization.
• Several participants expressed frustration with the proliferation of nearly identical note-taking apps, questioning the need for constant reinvention when core functionality remains consistent across existing projects.
• Significant debate persists regarding naming conventions, with many developers favoring generic, food-related names despite potential SEO and branding collisions, while others advocate for more distinctive identifiers.
• Legal and ethical compliance remains a critical friction point, particularly regarding two-party consent laws, with suggestions that apps should proactively notify all participants whenever a new person joins a call.
• Users are increasingly seeking deep customization, including the ability to define custom summarization prompts, trigger automation scripts (such as moving transcripts to Obsidian), and manually select specific AI models.
• The effectiveness of transcription often hinges on local hardware capabilities, leading to requests for better model selectors and offloading options for power-intensive tasks like summarization.
• Technical implementation details, such as the use of Unix sockets for local inter-process communication, are scrutinized by the community to ensure that local transcript access remains secure and authenticated.
• Many individuals have turned to these tools to manage meeting chaos, improve follow-ups, and mitigate the cognitive load of balancing active participation with manual note-taking.
The discourse reflects a broader community shift toward local-first AI tooling, driven by both privacy anxieties and the desire for granular control over personal data. While the market for meeting assistants is becoming saturated, developers are differentiating their projects through specific technical trade-offs, such as native UI performance versus cross-platform compatibility, or local-only processing versus API-driven model flexibility. Despite the recurring debates over naming and the ethics of silent recording, there is clear consensus that these tools have become essential productivity utilities for navigating the complexities of modern, information-dense work environments.