Speech Note

Notes you dictate, read aloud and translate, all processed on a Linux desktop with no network connection.

Instead of Otter.ai, cloud dictation

github.com/mkiol/dsnote

Screenshot from github.com/mkiol/dsnote, taken 2026-10-30, shown to help you recognise the tool. It belongs to its maker. Ask us to remove it.

Most dictation tools do one job and Speech Note does three. It turns speech into text, text into speech and one language into another, using engines such as Vosk, Whisper.cpp, Piper, Kokoro and Bergamot, and it says all of that processing happens entirely offline on your computer. It can insert dictated text into the active window, though on Wayland that needs ydotool, and it writes subtitles in SRT format.

Speech Note scores a 5 on lock-in, a 4 on footprint, self-direction and power and a 3 on longevity and delivery.

It runs on Linux desktops and Sailfish OS only, so Delivery is a 3. Models have to be downloaded before you use them, so do that before the flight. One developer makes it and asks for donations through Ko-fi or Liberapay, and the code is under the Mozilla Public Licence 2.0.

What stands out

• Speech to text, text to speech and translation, all processed offline
• Several engines, including Whisper.cpp, Piper and Bergamot
• Linux and Sailfish OS only

For comparison, we also reviewed Otter.ai (14/30) as a benchmark.