Software that gets out of the way.

We build tools that start instantly, do one thing well, and leave nothing running. The first one turns your voice into finished text.

No background processes when you're not dictating

One request per dictation, start to finish

No accounts no provider signup, no API key

No trackers in the app or on this site

One product so far

Takumi Voice

Dictation · macOS, Windows, Linux

Hold one key and say what you mean. The filler comes out, the punctuation goes in, and finished text lands in whatever you're already using. No window, no transcript to fix.

Read more

More Takumi tools are coming. They'll share one account and one licence page, and the same rule: nothing stays resident unless it's earned it.

Efficiency is a decision, not a cleanup pass

Small downloads, instant startup and nothing left running are choices you make at the start. You can't bolt them on later, so we don't try.

  1. 01

    Reliable by design

    It behaves the same way every time, fails in ways it can explain, and has fewer places to go wrong.

  2. 02

    Fast by architecture

    Speed comes from what we chose to build, not a tuning pass at the end. Hard to add later, easy to design in.

  3. 03

    Minimal overhead

    Every dependency, process and button has to earn its place. Most don't, so most aren't there.

  4. 04

    Craft in every detail

    The installer, the error messages, the update, the 2am edge case. That's the product, not packaging around it.

Capability does not require excess.

Takumi Voice is free for 2,000 words a week. No card, no account, nothing to cancel.