OpenWhispr

“Tried to pay for WhisperFlow. Indian card didn't work. So I just built it myself. OpenWhispr, voice-to-text that runs fully offline on your Mac. Hold Fn anywhere, speak, release, text gets pasted. WhisperKit + Qwen3 running locally on Metal. Open source, free forever: https://github.com/MrPrinceRawat/OpenWhispr If anyone here uses dictation tools or has been locked out of a paid tool because of payment issues, this is for you.”

“The app doesn’t allow me to download the local model. It gives me an error: The developers say in the description that the app is open source, but only the desktop version is open source. The mobile version is not open source, so there is no way to verify what data it sends, where it sends it, or whether any information is being sent to the cloud. Since the mobile app’s source code is not public, there is no way to independently confirm this.”

“seeing more Wispr Flow users say the quality has dropped lately, especially around accuracy, cleanup, and how much editing is needed after dictation. has anyone here moved to open source alternatives instead? I’ve seen people mention FluidVoice, FreeFlow, OpenWhispr, Handy, and a few Whisper based setups. main things I care about are accuracy, speed, formatting, cleanup, and whether it works smoothly across normal apps without needing constant fixing. for anyone who switched, what did you land o”

“this is a solid build, especially shipping fully offline on-device. one thing that could boost trust fast: add a tiny benchmark table in the README (model size, avg latency, WER on a short test set, RAM use) so people can compare it against macOS dictation and WhisperFlow quickly. also consider a push-to-talk key remap + per-app language/profile presets since that’s where dictation tools usually win daily usage. love that you built around a real payment pain point instead of just cloning a trend”

“Use Handy. I feel like I'm glazing it at this point (3rd post in a week about this). I found it better than OpenWhispr. It's lightweight, simple to use, and you can add hosted or local AI for post processing if you want.  I've been using it non stop since I found it.”

“i’d like to like this app, but…: I have a PhD. I was awarded that by an Ivy League school after a great deal of work. And one of the things I learned in earning that was do not rush your work to completion if it’s not complete.. So in trying to get this app operational, it literally does not go beyond step four of nine as it tells you on the upper left-hand corner of the app. There’s definitely nothing you can do after that for it to proceed. There’s no icon you can tap on. There’s no link you c”

“the transcription part is basically solved, whisper.cpp gets you most of the way and everyone shipping here has the same quality floor. so the "why this over openwhispr" question is fair and the answer can't really be accuracy. where these actually differ is the insertion layer, which field has focus, how it handles formatting commands, latency from stop-talking to text-appears. i built dictation into a desktop tool and that last 5% of UX ate more time than the whole model integration. that's th”

“Hello As the title says, I’ve been testing more than 10 of these apps mainly based on the apps comparison spreadsheet. I’m looking for a dictation app with the below requirements:- Free- allows local models (Parakeet)- Context aware (reads the screen and dictates accordingly, like asking to reply to an email)- Has edit mode for selected text The closest I found were: TypeWhisper but I don’t think it’s fully context aware and doesn’t have edit mode, FluidVoice which has the edit mode but I didn’t”