@nate_512Actually the wait is over for an open-source dictation tool that bundles local Whisper and NVIDIA Parakeet into one package. OpenWhispr sits at your cursor, so you can hit a hotkey, talk, and your words appear right where you're typing. You can also mark a section of text, give it a command, and it'll rewrite that selection without you lifting a finger. Privacy is the default here: everything runs on your own machine unless you deliberately hook up a cloud API. The feature list is stacked, too, on-device speaker diarization, voice fingerprinting, searching across your local notes, and full transcription for Zoom and Teams. There's even a built-in MCP server for connecting other agents. It's MIT-licensed, free, and positioned as a WisprFlow and Granola alternative, already pulling over 6,000 stars on GitHub.

@nate_512Actually the wait is over for an open-source dictation tool that bundles local Whisper and NVIDIA Parakeet into one package. OpenWhispr sits at your cursor, so you can hit a hotkey, talk, and your words appear right where you're typing. You can also mark a section of text, give it a command, and it'll rewrite that selection without you lifting a finger. Privacy is the default here: everything runs on your own machine unless you deliberately hook up a cloud API. The feature list is stacked, too, on-device speaker diarization, voice fingerprinting, searching across your local notes, and full transcription for Zoom and Teams. There's even a built-in MCP server for connecting other agents. It's MIT-licensed, free, and positioned as a WisprFlow and Granola alternative, already pulling over 6,000 stars on GitHub.