Getting Started

Welcome to Conjure! This page gets you from zero to dictating in under 5 minutes. For deeper dives on any topic, follow the links throughout.

Install

Beta

Conjure is in beta. Pre-built installers are coming soon. For now, build from source.

Clone the repository and run locally:

git clone https://github.com/Codename-11/conjure.git
cd conjure
pnpm install
pnpm run dev

Requires Node.js 20+, pnpm 10+, and the Rust toolchain. See the repo README for platform-specific dependencies. Conjure runs on Windows and Linux.

First launch

The onboarding wizard handles everything — just follow the steps:

1. Choose your mic — pick your input device and test it with a quick recording
2. Pick a transcription provider — local whisper (offline, private) or a cloud API key (faster)
3. Set your dictation hotkey — the key you'll hold while speaking
4. Try a practice dictation — the wizard walks you through it live

Recommended hotkey

Use Right Ctrl or Caps Lock — they're easy to reach and don't conflict with apps. Avoid Alt-based keys, which trigger menu bars on Windows.

Dictate!

Once setup is done, dictation works everywhere:

1. Click into any text field (editor, browser, chat, terminal — anywhere)
2. Hold your dictation hotkey
3. Speak naturally
4. Release — your text appears

That's it. The small pill overlay at the bottom of your screen shows the current state (recording → processing → done).

What to try next

You're up and running! Here's what to explore:

I want to...	Go to
Control punctuation by voice	Voice Commands — say "period", "new line", "question mark"
Change how my text sounds	Writing Styles — Polished, Email, Chat, Verbatim, and custom
Fix recurring transcription mistakes	Dictionary — auto-learn from corrections
Use AI to write, edit, or run commands	Agent Mode — voice-driven AI assistant
Hear text read back to me	Text-to-Speech — system or OpenAI voices
See my dictation statistics	Analytics page (in the sidebar)
Improve accuracy in noisy rooms	Audio Preprocessing — noise gate, filters
Use free cloud transcription	API Keys — Groq and Gemini have free tiers

Transfer settings between machines

Export your config (API keys, preferences, dictionary, tones, hotkeys) to use on another computer:

pnpm run export-config                      # Exports to ./conjure-config-backup/
pnpm run export-config -- --output ~/backup # Custom output path

Then copy the exported conjure.db file to the app data directory on your other machine:

• Windows: %APPDATA%\com.conjure.desktop\
• Linux: ~/.local/share/com.conjure.desktop/

Or use the in-app Settings → Data & Privacy → Export data / Import data dialogs for a ZIP-based transfer with selective category export.

Quick tips

• Voice commands work immediately — say "comma", "new paragraph", or "exclamation point" while dictating. No setup needed. See the full list.
• The pill overlay is interactive — hover to see your hotkey, click to start dictating, click during TTS to stop playback.
• Streaming mode shows text as you speak (real-time). Toggle it in the Styles section. Best with cloud providers.
• Everything is local by default — your audio and transcriptions stay on your machine. Cloud features are opt-in.
• Keyboard shortcuts — configure them all in Settings → Input → Hotkey shortcuts. See Keyboard Shortcuts for the full reference.

Need help?

• Click the ? icon in the sidebar for quick access to shortcuts and voice commands
• Check Troubleshooting for common issues
• Browse the FAQ for frequently asked questions
• Report issues on GitHub