Fonos – Open-source voice dictation for macOS, BYO key or local (github.com)

Fonos, an open-source voice dictation application for macOS and Linux, has been launched, allowing users to seamlessly capture and transcribe voice input using various speech-to-text (STT) and language model (LLM) providers. Inspired by technologies like Whisper Flow, Fonos operates on a local-first approach where users can either utilize their own API keys from major platforms such as OpenAI and Google or run the software entirely locally. This flexibility ensures privacy and control over captured information, making it easier to transcribe notes, generate meeting summaries, and engage with AI agents simply by holding a hotkey. The significance of Fonos lies in its ability to streamline workflows by transforming spoken language into structured notes and actionable items. With features like dual-channel audio capture for meetings, customizable transcription modes, and an integrated AI conversation panel, Fonos enhances productivity while keeping user data local. It supports real-time processing and customizable notebook exports to Markdown or JSON, catering to various workflows. Designed to operate with minimal distraction, Fonos uses global hotkeys for easy access to functions, thus providing a flexible and efficient solution for professionals in the AI/ML community seeking advanced dictation capabilities.