51 comments

[ 3.3 ms ] story [ 74.7 ms ] thread
Looks interesting. Why does it need a GUI at all?
The Parakeet V3 model is really great!
I don't get why Apple can't just integrate this and we're done with keyboards..
This looks great! What’s missing for me to switch from something like Wispr Flow is the ability to provide a dictionary for commonly mistaken words (name of your company, people, code libraries).
Use it daily. Looks and works great.
There's a slightly awkward naming overlap with an existing product.
Has anyone compared this with https://github.com/HeroTools/open-whispr already? From the description they seem very similar.

Handy first release was June 2025, OpenWhispr a month later. Handy has ~11k GitHub stars, OpenWhispr has ~730.

Creator of OpenWhispr here! Honoured to be compared to Handy!

I built OW because I was tired of paying for WisprFlow. I'd say it is more flexible by design: Whisper.cpp (CPU + GPU) for super fast local transcription, Parakeet in progress, local or cloud LLMs for cleanup (Qwen, Mistral, Gemini, Anthropic, OpenAI, Groq etc.), and bring-your-own API keys!

Handy is more streamlined for sure!

Would love any feedback :)

I just set this up today. I had Whispering app set up on my Windows computer, but it really wasn't working well on my Ubuntu computer that I just set up. I found Handy randomly. It was the last app I needed to go Linux full-time. Thank you!
A question because I'm not using speech-to-text, but find it intriguing (especially since it's now possible to do locally and for free).

How have your computing habits changed as a result of having this? When do you typically use this instead of typing on the keyboard?

Crashes on Tahoe 26.3 Betq 1 :(
Would be nice if the output can be piped directly into Claude Code.
On a M4 Macbook Air, there was enough lag to make it unusable for me. I hit the shortcut and start speaking but there was always a 1-2sec delay before it would actually start transcribing even if the icon was displayed.
I have dystonia which often stiffens my arms in a way that makes it impossible for me to type on a keyboard. TTS apps like SuperWhisper have proven to be very helpful for me in such situations. I am hoping to get a similar experience out of "Handy" (very apt maming from my perspective).

I do, however, wonder if there is a way all these TTS tools can get to the next level. The generated text should not be just a verbatim copy of what I just said, but depending on the context, it should elaborate. For example, if my cursor is actively inside an editor/IDE with some code, my coding-related verbal prompts should actually generate the right/desired code in that IDE.

Perhaps this is a bit of combining TTS with computer-use.

I made something called `ultraplan`. It's is a CLI tool that records multi-modal context (audio transcription via local Whisper, screenshots, clipboard content, etc.) into a timeline that AI agents like Claude Code can consume.

I have a claude skill `/record` that runs the CLI which starts a new recording. I debug, research, etc., then say "finito" (or choose your own stopword). It outputs a markdown file with your transcribed speech interleaved with screenshots and text that you copied. You can say other keywords like "marco" and it will take a screenshot hands-free.

When the session ends, claude reads the timeline (e.g. looks at screenshots) and gets to work.

I can clean it up and push to github if anyone would get use out of it.

Does anyone have a similar mobile application that works locally and is not too expensive? Mostly looking to transcribe voice messages sent over Signal which does not offer this OOTB
Did this thing (or open-whispr) work well with other languages than english ?
Love it. I had been searching for STT app for weeks. Every single app was either paid as a one off or had a monthly subscription. It felt a bit ridiculous having to pay when it’s all powered by such small models on the back end. So I decided to build my own. But then I found “Handy” and it’s been a really amazing partner for me. Super fast, super simple, doesn’t get in my way and it’s constantly updated. I just love it. Thanks a lot for making it! Thanks a lot

P.S. The post processing that you are talking about, wouldn’t it be awesome.

This is so handy, thank you very much. Good work!!
It’s incredibly fast on my MacBook m1 air and more accurate that the native speech to text.

The ui is well thought out, just the right amount of setting for my usage.

Incredible !

Btw, do you know what « discharging the model » does ? It’s set to never by default, tried to check if it has an impact on ram or cpu but it doesn’t seem to do anything.

the model is permanently loaded into ram for access speed. discharging it would unload it from ram and lead to longer start times
This looks and works great! A settings option to keep no recording history at all would be terrific.
It’s in the debug menu right now (ctrl/cmd+shift+d)
As a Mac user, am I missing something? macOS has Dictation built-in, when you short press F5 it should start transcribing your spoken words into text in real time. It even does non-English languages.
Besides being trash as others said, there’s a trade off with real time transcription word by word - there’s no opportunity for an AI to holistically correct/clean up the transcription
Is it deployed locally or does it send data to your servers?
Nice. I spent most of Christmas vibe coding with Google Antigravity with one hand while holding a sleeping baby in the other. MacOS built in dictation is OK, but struggles with technical language.
I’ve tried several, including this one, and I’ve settled on VoiceInk (local, one-time payment), and with Parakeet V3 it’s stunningly fast (near-instant) and accurate enough to talk to LLMs/code-agents, in the sense that the slight drop in accuracy relative to Whisper Turbo3 is immaterial since they can “read between the lines” anyway.

My regular cycle is to talk informally to the CLI agent and ask it to “say back to me what you understood”, and it almost always produces a nice clean and clear version. This simultaneously works as confirmation of its understanding and also as a sort of spec which likely helps keep the agent on track.

UPDATE - just tried handy with Parakeet v3, and it works really well too, so I'll use this instead of VoiceInk for a few days. I just also discovered that turning on the "debug" UI with Cmd-shift-D shows additional options like post processing and appending trailing space.

Explain to me why a speech-to-text app has 50% of its code in typescript...?