First-time setup

After Install & permissions, follow this workflow to try dictation. Figures are operation diagrams showing configuration order and data paths; the current app determines actual appearance.

Overview

Step	Where	Goal
1	Menu bar → Settings	Open configuration
2	General	Check recording permissions and input device
3	Voice models	Choose an engine; install resources or enter credentials
4	Shortcuts	Check hotkeys and trigger modes
5	Text processing / General	Check processing and output action
6	Notes or another text field	Test real recording and insertion

1. Open Settings

Choose VoiceHotkey menu bar icon → Settings. The main sidebar lists General, Shortcuts, Voice models, Text processing, Voice commands, File transcription, History; License and About appear at the bottom.

Open Settings and navigate (operation diagram)

You can also open voicehotkey://tab/settings; see Deeplinks.

2. Check recording readiness

New installations default to real ASR. You do not need to turn Demo off, and current General settings has no Demo switch.

Under Settings → General, check Microphone, Accessibility, and your input device. Address any missing-permission or model warnings in the menu bar. If a legacy configuration or test environment still shows Demo status, it means demo text is used; it does not establish that all cloud requests are disabled. See BYOK & data flow.

Real recording by default and readiness checks (operation diagram)

3. Choose a voice model

1. Open Settings → Voice models and select an available engine. Apple can be a first option; on-device models need their model and runtime resources installed first.
2. For cloud ASR, enter the required credentials, model, and endpoint. API keys are stored in Keychain, and requests go directly to that provider.
3. Check for missing-credential or resource warnings. See Voice models and Plans for system / hardware requirements and Free limits.

Choose recognition and resources / credentials (operation diagram)

Local Only Mode

Public ASR endpoints are blocked while it is enabled. Use Apple, an installed local model, or a loopback service. It restricts ASR / LLM and pauses updates; license requests can still use the network. See Local Only Mode.

4. Confirm shortcuts

Open Settings → Shortcuts. Voice typing defaults to Right Command, commands to Middle mouse button, and translation to Right Control. Direct Voice Typing is unbound by default.

The default trigger mode supports tap to start / tap again to finish, or hold to speak / release to finish. Rebind if another app conflicts. Quick Send and Esc cancellation controls are on Shortcuts; audio feedback is on General. See Shortcuts & recording.

Shortcuts and recording triggers (operation diagram)

5. Text processing and output

Start with simple settings and add processing after dictation works:

Option	Current default / location
Strip trailing period	On; Text processing
Smart spacing	Off; Text processing, enable if useful
Word replacements	Text processing; add for recurring errors
LLM polish	Off; cloud needs credentials, local needs installed models; adaptive polish requires Trial / Pro
Output action	General → Output action; defaults to insert at caret
Translation	Select engine and target in Text processing; Apple native requires macOS 26+, or configure an LLM

Smart paragraphs are not a current feature. History-only mode is an internal setting and not offered in the current output menu; that mode saves to local history without insertion or automatic copying.

Processing and output choices (operation diagram)

6. End-to-end test

1. Open Notes or TextEdit and click an editable text field so the caret blinks.
2. Start recording with the voice typing hotkey and speak a short sentence.
3. Tap again or release the held shortcut to finish.
4. Wait for recognition and enabled processing, then check for text in the focused field.
5. Check History for the result and status. Without an editable focus, the default can fall back to the clipboard; behavior depends on the output action and related settings.

Complete a dictation session (operation diagram)

If it fails

Symptom	Go to
No hotkey response	Troubleshooting · hotkey
Recording but no text	Troubleshooting · no text
Permissions keep dropping	Troubleshooting · permissions

After setup

Word replacementsFix brand names and recurring errors.

Local Only ModeRestrict ASR / LLM and pause updates; understand the license network exception.

Voice commandsCreate phrases and configure supported actions and an optional intent model.