Sam W. made me stop typing prompts
Last Friday at the 2026 SSW Brisbane brainstorming, I got to work with Sam W.
If you have worked with Sam, you know the vibe. He is chill, practical, and usually 2 steps ahead with AI workflows.
I thought my prompting setup was fine. It was not.
My old workflow looked like this:
- Open ChatGPT
- Click the transcribe button
- Speak
- Copy text
- Paste into OpenCode CLI or a Claude web app (I am cheap, no Claude sub)
- Repeat until my soul left my body
Or even worse, type the prompt manually 🤢 (that was Sam's face when he saw me typing).
Then Sam showed me Handy, a free and open source speech-to-text app.
That changed everything.
handy.computer
I downloaded the app from handy.computer.
On 1st launch, Handy asked me to choose a local model. Sam told me to pick OpenAI Whisper Turbo. I picked it. No debate 🫡 (Sir Yes Sir).
Once the local model finished downloading, I gave Handy two permissions:
- Microphone access
- Accessibility access (so it can paste text automatically)
Done.
My new prompting loop
On Mac, I press Option + Space.
I speak.
Handy transcribes and pastes directly where my cursor is.
Then I hit Enter and keep going.
On Windows, the default shortcut is Ctrl + Space.
That is it.
No copy-paste dance. No app switching. No transcription API cost.
Just hold Option + Space, speak, enter, repeat.
Why I am sticking with Handy
This tiny workflow change has made a real difference for me.
I can prompt faster, stay in flow, and skip the ChatGPT copy-paste loop because everything runs locally.
If you are still transcribing using ChatGPT web app and pasting into another, try Handy:
Your keyboard deserves a break.
Comments
To edit a comment, click its timestamp. It will open the GitHub Discussion where you can edit your comment.