TongueType for macOS - Local dictation for macOS without the subscription
TongueType is a macOS voice dictation app powered by Whisper AI running locally on Apple Silicon. No cloud, no accounts, no subscriptions. Hold a key, speak, release. Your words appear. Supports 12 languages and audio file transcription.
TongueType gives you a configurable press-to-talk hotkey, audio and video file transcription, and configurable post-processing rules.
It's customizable and fun (try Rainbow Mode!) and it's built to be the fastest dictation workflow possible.
Replies
TongueType for macOS
local-only and no subscription, rare combo! how do you handle the punctuation problem. say "comma" and it adds punctuation? how would "period" be decided as word or punctuation? congrats on your launch!
TongueType for macOS
@hiyamojo I've gotten used to not using it, to be honest. But you can go to Settings > Postprocessing and experiment with replacements there! A friend of mine added, e.g. "smiley face emoji"' => 😊 and similar.
For me, the thing that breaks the dictation habit is the cleanup tax after the words appear.
Dictation is great for raw thought, but different surfaces need different levels of cleanup: an LLM prompt can stay loose, a commit message needs precision, and an email needs just enough polish without losing the spoken cadence. The post-processing rules feel like the right place to solve that.
One thing I’d want is per-context presets plus a quick raw transcript / cleaned text comparison. That would make it easier to trust the tool because you can see whether it’s preserving the thought or silently over-smoothing it.
TongueType for macOS
@jim_jeffers Thanks for the feedback! What would per-context presets look like and how would they operate?
I’d think of them as small rewrite policies tied to the destination app, with a default fallback.
For example: Messages/Slack = keep it casual, trim false starts, don’t over-punctuate. Email = clean grammar and paragraph breaks, but keep my wording. Git commits = short imperative, no hedging. LLM prompt = preserve messy detail, only fix transcription errors.
Operationally, the lightweight version could be app-based rules plus a quick chooser/hotkey when the app guess is wrong. The raw/cleaned diff matters because if the preset silently makes me sound too polished, I’ll stop trusting it.
TongueType for macOS
@jim_jeffers I'll think about this. I wonder if it can be done with a modifier key or spoken word, but that's a neat idea!
@claviska If you want a tiny gut-check, the risky case is: raw dictation → cleaned output where the thought is technically clearer but no longer sounds spoken by me.
Even one anonymized raw/cleaned pair from your own usage would reveal whether the preset should be “remove friction” or “rewrite into prose.”
This needs post AI rewrite in order for me to use it. Like what Typeless.com is doing.
TongueType for macOS
@patrickpetcejj I can look into this, but I worry that a local pass on an LLM would slow things down. Are you mostly looking to remove things like "um" and cleaning up false starts?
@claviska if I speak something, I don't have to worry exactly what I'm saying. I can just ramble, and it will get rid of the ums and everything.
Check out how Typeless does it. If you could make it work how they do it, you would do so well with your product.
Documentation.AI
Congrats. Where do I see pro and free features list?
TongueType for macOS
Thanks! @roopreddy free includes every feature, but capped at 30 minutes of dictation per month and 10 seconds for file transcriptions. There's a Free vs. Pro page with more details.