Quickstart Guide

Get dictating in under 2 minutes.

1. Install

Run the installer. It downloads the latest binary and offers to install ydotool for text injection.

curl -fsSL https://get.omniscribe.dev/install.sh | sh

Requires ydotool for text injection. The installer offers to install it if missing.

2. Start the daemon

The daemon runs in the background, managing audio capture, STT, and text injection.

omniscribe-core daemon

3. TUI dashboard (optional)

Sign in, upgrade, view usage, and switch modes — all from the terminal.

omniscribe-core tui

4. Dictate

1. Click any text field (browser, terminal, IDE, Slack — anything)
2. Press F9 to start recording
3. Speak clearly
4. Press F9 again to stop and inject the transcribed text

Text appears wherever your cursor is. No copy-paste needed.

5. Sign in

Open the TUI and sign in with Google, GitHub, Microsoft, or email. Free tier: 500 words per 5-hour window. Pro ($5/mo): unlimited words, works fully offline.

6. Polish modes

OmniScribe can clean and format your dictation before injecting it. Each mode has its own hotkey:

All polish processing is local. No text leaves your machine unless you configure your own LLM keys.

7. Configuration

Edit ~/.config/omniscribe/config.toml to customize hotkeys, microphone, STT model, VAD threshold, and more.

[hotkey]
combo = "F9"

[stt]
model_override = "ggml-base"

[mic]
device = "default"

8. Uninstall

omniscribe-core clear && rm ~/.local/bin/omniscribe-core