meetily: Transcribe and Summarize Meetings Locally

Overview

Meetily is a desktop application for capturing meeting audio, producing transcripts, and generating summaries. Its local transcription and storage aim to keep recordings and transcripts on the user's machine, while summaries can use Ollama or configured remote AI providers.

It is a fit for individuals and organizations that need meeting notes but want to control where sensitive audio and text are processed. It supports macOS, Windows, and Linux, though installation and acceleration options vary by platform.

Key Features

• Live transcription with Whisper or Parakeet models, processed locally.
• Meeting summaries with Ollama or supported providers including Claude, Groq, OpenRouter, and OpenAI-compatible endpoints.
• Microphone and system audio capture at the same time.
• Import audio files for transcription or re-transcription with a different model or language.
• Local storage for recordings, transcripts, and transcription models.
• GPU acceleration options vary by platform and build, including Metal/CoreML on macOS packages and Vulkan on Windows packages.

Use Cases

• Professionals who need searchable transcripts and summaries of recurring meetings without routinely uploading recordings to a transcription service.
• Legal, healthcare, or compliance-conscious users who need greater control over sensitive meeting data and can verify the application's processing and storage setup for their requirements.
• Individuals working offline or with local AI models who want meeting notes without relying on a cloud transcription API.
• Developers or teams with custom AI infrastructure who want to route summaries through an OpenAI-compatible endpoint.

Getting Started

Download the macOS or Windows installer from the latest releases. Linux users can build from source; see the README and its build guides for platform-specific steps.

Alternatives

• Mityu: Mityu is a fully local Rust and Tauri meeting assistant, while Meetily also supports configurable AI providers.
• Speakr: Speakr is a self-hosted web app for transcribing audio recordings into searchable notes, rather than a desktop meeting recorder.

Considerations

• “Local” applies to transcription and stored meeting data, but summaries may be sent to a remote provider if one is configured. Use Ollama for a local summarization path and check provider settings before processing sensitive material.
• Hardware acceleration and compatibility depend on the operating system and build. The packaged Windows installer requires an AVX2-capable x64 CPU; CUDA requires a suitably configured source build.
• Linux installation is source-based according to the README, so it requires development tools and additional setup.
• The README distinguishes the free Community Edition from a separate PRO offering. Check the project's release notes and documentation for current feature availability.