MurmurFlow is a voice dictation app for macOS. It turns spoken input into polished, punctuated prose in under a second and is described as 100% on-device, offline, and fully local on a Mac.
Its current features include an on-device ASR engine powered by FluidAudio and WhisperKit, which produces local transcription. It also includes intelligent text cleanup using local Apple Foundation Models for filler-word removal, capitalization, and punctuation. Other listed functions are deep app context that tracks active windows, app identifiers, and target textboxes; an auto-learning dictionary for vocabulary and correction mapping; voice command mode for reformatting by speech; adaptive audio resilience for mic-gain leveling and trailing-silence trimming; tone and style matching; context-aware pasting; custom polish prompts; native text snippets; a voice notes hub; and visual OCR extraction. The page also says it can import a dictionary, corrections, and context from Wispr Flow locally.
MurmurFlow is presented for people who want private dictation on a Mac, especially those comparing it with cloud dictation tools that upload voice audio, screen contents, active apps, passwords, identity, and telemetry to remote servers. The site says it keeps everything isolated on the user’s Mac and that no internet connection is established during processing.
Delivery is limited here to macOS. The page refers to an early access waitlist and says the product is free during Early Access, with no subscriptions. It also states that there are no accounts or databases, no cookies, trackers, or analytics, and that transcribed data is not stored outside the local filesystem.
MurmurFlow is a Voice, TTS & speech project. It focuses on transcribing speech to text on macOS devices without compromising privacy or requiring cloud services. MurmurFlow is a consumer product aimed at macOS users needing voice dictation. MurmurFlow costs nothing to use. It ships for the web and macOS.
MurmurFlow first shipped in 2024. Among its 5 catalogued features are on-device transcription, offline mode, and voice command mode. MurmurFlow is currently in beta.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do