Developer releases open-source AI meeting recorder for Mac

By Splitfeed EditorialLast updated Sep 30, 2026, 18:00 UTC

On September 30, 2026, a London-based entrepreneur named Uygar posted a Show HN submission to Hacker News introducing Parrot, an open-source meeting recorder built for macOS. Uygar describes himself as someone who runs his businesses through daily calls with customers, partners, and suppliers, and who built the tool over several months with assistance from the AI coding assistant Claude to solve his own memory and note-taking problem. The app records system audio locally, requires no account, and makes no external API calls when using local models, meaning all recorded data stays on the user's machine. It also includes a co-pilot feature for meetings, though the full description of that feature was not available in the submission as published.

The story matters less as a product announcement and more as an illustration of a pattern that has become increasingly common: individual developers using AI assistants to build functional, privacy-focused software tools that would previously have required a small team or significant funding. The privacy angle is particularly relevant in the meeting recorder category, where cloud-based competitors such as Otter.ai or Fireflies have drawn criticism for uploading sensitive business conversations to external servers. A local-only, open-source alternative shifts the trust model entirely to the user's own hardware and the transparency of the code itself.

None of the three camps have posted reactions to this specific story yet. The Pro-AI camp would typically point to Parrot as a clear example of AI democratising software development, allowing a solo entrepreneur to ship a useful product in months rather than years. The Anti-AI camp would be expected to raise questions about the quality and auditability of AI-assisted code, and whether widespread use of tools like Claude to generate software creates subtle reliability or security risks that solo developers may not catch. The Middle Ground camp would likely focus on the privacy benefits of local processing while asking for more transparency about what Claude contributed and how the codebase has been reviewed.

The argument is most likely to sharpen once Parrot attracts independent code review from the Hacker News community, which routinely scrutinises Show HN submissions in the comments. Any audit of the codebase, community fork activity on the project repository, or follow-up from Uygar addressing the co-pilot feature details would give a clearer picture of how robust the tool actually is.