Windows Speech Recognition Registry Key Location

This article details the exact Windows Registry keys used to store configuration parameters, profiles, and settings for Windows speech-to-text recognition services. Depending on whether you are configuring modern Windows 10/11 dictation services or legacy Windows Speech Recognition (SAPI), the settings are stored across distinct locations in the Current User and Local Machine hives.

Modern Speech-to-Text Settings (Windows 10 & 11)

Modern Windows dictation and cloud-assisted speech recognition engines (OneCore) store their core configuration under:

• User Configuration:
  HKEY_CURRENT_USER\Software\Microsoft\Speech_OneCore
• System-Wide Configuration:
  HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech_OneCore

Important OneCore Subkeys:

• Recognizers:
  HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech_OneCore\Recognizers\Tokens
  Contains installed speech recognition engines, language packs, and engine-specific parameters.
• User Preferences & Calibration:
  HKEY_CURRENT_USER\Software\Microsoft\Speech_OneCore\Preferences
  Stores user-specific recognition settings, audio input preferences, and adaptation parameters.
• Online Speech Recognition Policy:
  HKEY_CURRENT_USER\Software\Microsoft\Speech_OneCore\Settings\OnlineSpeechPrivacy
  Contains the HasAccepted DWORD value (where 1 enables cloud-based speech recognition and 0 limits recognition to on-device processing).

Legacy Windows Speech Recognition & SAPI Settings

Classic Windows Speech Recognition tools, third-party SAPI-compliant software, and legacy desktop accessibility features store parameters under:

• User Configuration:
  HKEY_CURRENT_USER\Software\Microsoft\Speech
• System-Wide Configuration:
  HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech

Important SAPI Subkeys:

• Installed Recognizers:
  HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech\Recognizers\Tokens
  Lists all legacy speech recognition engines registered with the operating system.
• User Profiles & Training Data:
  HKEY_CURRENT_USER\Software\Microsoft\Speech\CurrentUserProfile
  References the active voice profile GUID used for user voice training and acoustic adaptation.
• Audio Input Endpoints:
  HKEY_CURRENT_USER\Software\Microsoft\Speech\AudioInput
  Maintains the mapping to the default audio capture device used for legacy speech services.

How to Access These Keys

1. Press Win + R, type regedit, and press Enter.
2. Copy any of the paths above and paste them into the address bar at the top of the Registry Editor.
3. Press Enter to navigate directly to the target configuration parameters.