Vendor page
voiceio, dictation that learns you (home page)
voiceio (Hugo0), 2026
Read the originalRead on October 22, 2026
- What it is
- The home page of voiceio, a local voice dictation tool for Linux that runs Whisper on the user's computer, inserts text into the focused app and can fine-tune its model on the user's own recordings. The page links to its agent runbook and its GitHub repository.
- How much weight it carries
- The maker's own home page, read 2026-10-22. The project is open source under the MIT license and is run by one developer. The page describes Linux desktops; it does not describe Windows or macOS support.
Facts taken from this source
- The voiceio home page says there is no account, no server and no upload, so the user's recordings stay on the user's computer.
“No account, no server, no upload: it works offline, and there is no telemetry.” (How it works, step 02)
- The voiceio home page says the model runs on the user's computer, with Whisper decoding speech on the user's own CPU.
“The model runs on your computer. Whisper decodes your speech on your own CPU.” (How it works, step 02)
- The voiceio home page says text streams into the focused app, such as an editor, browser or terminal, after the user presses the hotkey and talks.
“Press your hotkey and talk. Text streams into the focused app” (How it works, step 01)
- The voiceio home page says its vocabulary.txt file holds the user's names and terms, ranked by how often they say them, and that the user can edit it.
“Your names and terms, ranked by how often you say them” (Privacy section, file table)
- The voiceio home page says the tool works offline, with no account, server or upload, and no telemetry.
“No account, no server, no upload: it works offline, and there is no telemetry.” (How it works, step 02)