TypeWhisper

@ahhhhfs A lot of dictation tools feel similar until longer-form writing is involved. TypeWhisper is stronger on engine choice, cleanup, and workflow tuning: typewhisper.com

Mac dictation inaccurate? Cloud tools make you worry about privacy? Input 0: Open-source Local Voice Input Tool for macOS Runs speech recognition on-device with offline models. Flexibly connects to your custom LLM to instantly fix grammar and polish text.

TypeWhisper 1.2 is out: faster startup, lower idle CPU, better AirPods/Bluetooth recovery, better short-clip handling, and a new Minimal indicator. Free, open source macOS dictation with local or cloud engines. github.com/TypeWhisper/ty…

@vad3rt3sla The main difference between dictation tools usually shows up after the first transcript. TypeWhisper puts more emphasis on cleanup, engine choice, and workflow control: typewhisper.com

Nobody wants Apple CarPlay in a Tesla

@csaba_kissi A lot of dictation tools feel similar until longer-form writing is involved. TypeWhisper is stronger on engine choice, cleanup, and workflow tuning: typewhisper.com

Share your website/project, guys👇

@uncledoomer @justalexoki I had the same frustration. The problem is not that voice input is impossible, it is that people get locked into one speech stack with no tuning and no context switching. TypeWhisper is my attempt to fix that on Mac with multiple engines, optional local processing, and per-app

@justalexoki it's so fucking embarrassing that apple let enshittification make speech to text worse over time. i can remember clearly that it was better 7-8 years ago. there are very obvious ways it can be made better and they just dont care

half my moots are retweeting these guys. are they friends? are they all paid? what's going on. there's no way they're all using this

@thatpeterdev Thank you, this is exactly the workflow I care about. TypeWhisper is built around fast Mac dictation with local or cloud engine choice, so getting that “near instant and accurate” reaction feels very right.

Just tried @Type_Whisper and honestly shookth. Near instant and so accurate 😭

@aiwithmayank That shortcut is great for quick input. The gap usually shows up later in cleanup and workflow. TypeWhisper focuses more on that side: typewhisper.com

5/ Disable the DiagTrack Service (the engine running all of this) DiagTrack is the service officially called "Connected User Experiences and Telemetry." It's the actual background process that packages your data and ships it to Microsoft. Turning off the Settings toggles tells it to send less. Disabling the service stops it from running at all. → Press Win + R → type services.msc → hit Enter → Scroll to "Connected User Experiences and Telemetry" → Double-click it → set Startup type to "Disabled" → Click Stop, then OK Note: Windows updates may re-enable this. Check back after major updates.

@skill_afrika_ That shortcut is great for quick input. The gap usually shows up later in cleanup and workflow. TypeWhisper focuses more on that side: typewhisper.com

Windows key + H Tired of typing? Pressing Windows + H instantly activates Windows' built-in voice typing feature. I use this on lazy days when my fingers aren't in the mood 😅 Believe me, you wouldn't want to type after this.👀

@amical_app A lot of dictation tools feel similar until longer-form writing is involved. TypeWhisper is stronger on engine choice, cleanup, and workflow tuning: typewhisper.com

your dictation app shouldn't need you to think about formatting. @amical_app does it automatically. professional in gmail. casual in slack. code-ready in cursor. 🗣️ just talk! comment "VOICE" and if you've ever rage-quit because of slow typing, we'll dm you the fix 😉

@Perp338 That shortcut is great for quick input. The gap usually shows up later in cleanup and workflow. TypeWhisper focuses more on that side: typewhisper.com

i just found out that pressing windows keys + H lets you use voice typing on your computer so you can talk instead of typing -i wish i knew this earlier

@sidpanj @pranavmitl Mostly restrictions, yeah. macOS is much more open for system-wide dictation workflows. On iOS you are dealing with tighter sandboxing, keyboard-extension limits, and less control over cross-app text handling.

@Type_Whisper @pranavmitl No worries - dm me when it’s ready to test. Out of interest, is there anything that makes it more technically difficult than the Mac app? Or is it a prioritization thing?

Apple has been asleep at the wheel on OS-native dictation for so long. Aqua feels like a great acquisition candidate to change that. Wispr Flow is too expensive

@abustamante @WisprFlow Yeah, that kind of inconsistency is brutal in speech-to-text. Even small quality swings feel huge because you stop thinking about the content and start babysitting corrections instead. That trust gap is a big part of what we care about with TypeWhisper: typewhisper.com

I'm almost convinced @WisprFlow A/B tests live a smaller model every now and then to see people's behavior/corrections. I can tell you already guys: it doesn't work very well.

@abhinav_sidhu @aquavoice @gritfollow @WisprFlow If the India pricing is really 3x, then that is a fair callout. Price claims only work if they hold up by region. Otherwise the better comparison is workflow, accuracy, and how much control you get.

@aquavoice @gritfollow @WisprFlow Guys, you are no way cheaper to Wisprflow!! Do check App Store prices and make such claims … for India App Store, your price is 3X … yes 3X!!!

Aqua Voice is now live for iOS. It's a premium voice keyboard for every app on your phone.

@srikruth7 Yeah, this is usually not about raw model quality alone. For coding voice mode, the real gap is project context: custom vocabulary, file names, symbols, and cleanup. Without that, codebase-specific terms get butchered fast.

If Mythos is so good why doesn’t Anthropic use it to fix CC’s voice mode? It has zero awareness of the codebase. It’s currently very basic (deepgram nova v3 with word boosting) and butchers most codebase-specific words. Even WisprFlow is better than this

@sidpanj @pranavmitl Sorry, not in a good state yet on our side. iOS is planned, but I’d rather be honest than send you to something half-baked.

@Type_Whisper @pranavmitl How can I test iOS?

@rohanpdofficial Since you built it yourself, TypeWhisper might be interesting for the cost/control angle: typewhisper.com. It's open source, no subscription, and gives you local or cloud engine choice.

I built a Wispr Flow alternative — but with way more than just dictation. $15. Once. Forever. No subscription. Wispr Flow gives you speech-to-text for $15/month. Dictar gives you a full voice productivity suite for $15. Once. Dictate : real-time streaming into any app, auto-punctuation, 99 languages Read Aloud : select any text, 51 natural voices, adjustable speed AI Rewrite : say "make this professional" and on-device AI rewrites it Snippets : type /greet, full paragraph expands instantly Custom modes for Email, Slack, Notes. Works in Notion, Gmail, VS Code, Figma — anywhere you type. No cloud. Works offline. Your voice never leaves your Mac. dictar.app

@Nick__Losee @problogger For speech-to-text workflows, maybe check TypeWhisper: typewhisper.com. It gives you local or cloud engine choice and a more flexible dictation setup if you want something beyond the default built-in tools.

@problogger Hmm I'm realizing my church has a mixed bag on their YT of Livestreams with and without transcript. I think i can get all the audio as mp3 but would you recommend any opensource Speech to Text models? Think i saw Microsoft released one today...

Over the past few weeks I've been transcribing 20 years of our church's sermons and bringing them into my Obsidian Vault where they've been tagged and classified based on topic, scriptures, speaker etc.

@sheherenow_ @tenobrus If cost and flexibility are the issue, maybe check TypeWhisper: typewhisper.com. It gives you local or cloud engine choice, plus a more configurable dictation workflow than the usual built-in options.

@tenobrus what do you prefer about it? specifics or just better? I've been wanting to replace Wispr Flow bc it's criminally expensive for what it is but almost every alternative is desktop-only

holy shit just replaced whispr flow with this and it's instantly a 10x better experience

@gkeaton23 If local matters, maybe check TypeWhisper too: typewhisper.com. It gives you engine choice, works beyond the basic built-in dictation flow, and is a better fit if you want more control over the workflow.

Who has a Wispr Flow alternative that is free? I am really enjoying the great transcription but don't want to pay $15/mo. Open source and local would be awesome too I have tried the built in Apple dictation but its no where close