---
title: "Yubi: A voice for your Mac that never leaves it | Carlos Lastres"
description: "A voice assistant in the Mac menu bar. Hold a key, talk, let go. Speech to text on device, a yes before anything risky, and a character who only speaks recorded lines. Free to try, direct download."
url: https://lastres.ai/work/yubi/
lang: en
---

Yubi 2026 Own product / macOS and AI

# A voice for your Mac that never leaves it

Yubi (ゆび) lives in the menu bar. Hold the right Option key and talk, and the words land where the cursor is. Hold the right Command key and ask, and she reads the screen and clicks for you, after you say yes. Speech to text runs on the Mac. The app is 6 MB. I designed her and the site, wrote the Swift app, the Cloudflare Worker behind it and the billing, and shipped it as a direct download in September 2026

![The Yubi homepage: talk to your Mac, she does the typing](https://lastres.ai/media/work/yubi/01-hero.webp)

\[ What it is \]

0

Seconds of audio uploaded. Speech to text runs on the Mac

6 MB

The whole app, notarized, direct download

2

Keys. Right Option dictates, right Command asks

- **Role:** Founder, Product Designer and Engineer
- **Timeline:** September to October 2026
- **Scope:** Product design, Character and art direction, macOS engineering, Cloudflare Worker and metering, Billing

## The challenge

The assistant I was using weighed 785 MB and sent my voice to the cloud

I wanted to talk to my Mac the way I talk to a colleague across the desk: say it once, have it typed in the app I am already in, and when I am lost, have someone point at the button. The tool I used for this shipped cloud speech recognition, screen recording, two analytics SDKs and a bundled command line agent, 785 MB in all. It guessed whether I wanted dictation or an action, and guessed wrong often enough that I stopped trusting it. And it spoke back in a synthesized voice that made every answer feel like a phone tree.

## What I did

Two keys, one yes, and nothing synthesized

Right Option types what you say. Right Command asks her to help with the screen. Two keys instead of one guess, because a tool that has to infer your intent is a tool you have to watch. Speech recognition runs on the Mac with Apple's own engine, so no audio ever leaves it. The model call goes through a small Cloudflare Worker that holds the key, meters each task and streams the answer back; the app never carries a credential. Before anything risky, sending a message, pressing Return, opening a URL, she shows exactly what will happen and waits for your yes. And she never speaks generated text. Her voice is a set of recorded Japanese lines, so she sounds like one person, every time.

The one bit manga art is a design decision, not a style. A black and white sprite stays sharp at 48 pixels in the menu bar and at full size on the site, and it reads as drawn, not rendered. The same character now sits on a desk display and on a sheet of LINE stickers, which is the kind of thing that happens when you design the person before the product.

## Outcomes

Hold a key, talk, let go. The words land where the cursor is, and nothing leaves the Mac

- Dictation into any app, unlimited and free, with speech recognition on the Mac. No audio is uploaded.
- Right Command reads the screen and clicks, types or opens things for you, and asks before anything it cannot undo.
- Sign in is a magic link. Free gives 10 tasks a month, Personal ¥1,480 for 60, Pro ¥2,980 for 150. Stripe inside the app.
- The app is 6 MB, notarized, and runs in English and Japanese.
- Character, site, Swift app, Worker, metering and billing, all designed, built and run by me.

Live at yubichan.ai, versions 0.6.0 to 0.8.1 shipped between September and October 2026.

[Working on something like this? Tell me about it.](https://lastres.ai/work-with-me/)

\[ What it actually looks like \]

![The Yubi listening panel with a Japanese sentence being transcribed and the status On this Mac](https://lastres.ai/media/work/yubi/02-hud.webp)

*The listening panel. The footer says where the audio is: on this Mac, always*

![The Yubi ask card showing a request to open a Chrome tab, what will be sent, and Cancel and Send buttons](https://lastres.ai/media/work/yubi/03-ask.webp)

*The yes gate. Before a task starts she shows the request, the app, the byte count and the model, and waits*

![The seven things to say to your Mac section, with a playable Slack scene](https://lastres.ai/media/work/yubi/04-uses.webp)

*Seven use cases you can play on the site. Each one is a key, a sentence, and the app it lands in*

![Her face in close up next to the heading Your voice never leaves this Mac and four facts](https://lastres.ai/media/work/yubi/05-voice.webp)

*The privacy page is four numbers. Zero seconds uploaded, 6 MB, no screen recording, no analytics in the app*

![The pricing section: Free, Personal and Pro](https://lastres.ai/media/work/yubi/06-price.webp)

*Talking is free. Help is cheap. Dictation never counts, only tasks do*

![The Japanese homepage with the name ゆび and the same portrait](https://lastres.ai/media/work/yubi/07-hero-ja.webp)

*Japanese is the first language, not a translation. The name is the word for finger*

[Visit Yubi ↗](https://yubichan.ai/en/) [Start something like this](https://lastres.ai/contact/)

[← All work](https://lastres.ai/work/)

---

Source: https://lastres.ai/work/yubi/
Site index for agents: https://lastres.ai/llms.txt
