Beta

Windows is currently a public beta. Profile matching already works well for testing, but polish and edge-case handling are still evolving.

Profiles

Configure per-application transcription settings that activate automatically.

How Profiles Work

When you start dictating, TypeWhisper matches the active application's process name against your profiles. For browser-based apps, it also matches URL patterns. If a match is found, the profile's overrides are applied automatically. The active profile name is shown as a badge in the recording overlay.

Matching

Profiles match using the process name (e.g., chrome.exe, outlook.exe) and optionally a URL pattern for browser tabs.

URL Pattern Syntax

• * matches any sequence of characters within a single segment
• *.github.com matches all GitHub subdomains (e.g., gist.github.com)
• docs.google.com/document/* matches all Google Docs documents
• *slack.com* matches any URL containing slack.com

Finding an App's Process Name

Open Task Manager (Ctrl+Shift+Esc), go to the Details tab, and find the application. The "Name" column shows the process name you need (e.g., Teams.exe, WINWORD.EXE).

Priority

When multiple profiles could match, the most specific one wins:

Process Name + URL Pattern (most specific)
URL Pattern only (cross-browser profiles)
Process Name only
Default settings (fallback)

Creating a Profile

Go to Settings > Profiles and click "Add Profile". Assign one or more applications, then configure your overrides:

• Language - Override the transcription language
• Task - Transcribe or Translate
• Engine - Parakeet TDT 0.6B or Canary 180M Flash
• Whisper Mode - Boosted microphone gain

Example Setups

Chrome - GitHub

Process: chrome.exe, URL: *.github.com, Language: English

Outlook - German

Process: outlook.exe, Language: German, Engine: Canary 180M Flash

Terminal - Whisper Mode

Process: WindowsTerminal.exe, Whisper Mode: Always on

Settings Location

Profile configuration is stored in %LOCALAPPDATA%\TypeWhisper\settings.json.