I built Dictly to speed up how I talk to coding agents - now it’s my daily tool

I built Dictly because I spend a lot of time talking to coding agents, and typing out my thoughts was always slowing me down.

Apple’s built-in dictation is good - fast, accurate enough, and private.

However, I wanted something tuned to my own workflow.

So I started experimenting with Apple’s Speech Recognition framework, just to see how far I could push it.

The first goal was simple: make dictation that feels instant and local, but still flexible.

From there, I added the things I always wished existed:

• 🗣️ Custom dictionary profiles — teach it names, brands, or syntax it keeps missing.

• ⚙️ AI post-processing pipelines — still on-device, but with cleanup and style tweaks for different contexts.

• ⚡ Quick Capture View — the killer part for me. A small overlay I can summon with a hotkey, dictate into, and paste directly into whatever app I’m using.

It’s not revolutionary — it’s just a solid, private dictation tool built the way I wanted it to work: fast, accurate, and local.

What started as a small helper for my own workflow turned into something I now use daily, and finally decided to share.

I’m curious — have you ever built a tool just to fix an annoyance in your own workflow that ended up becoming a “real” product?