FiveM-VoiceControls

FiveM Voice Controls

Control FiveM using your voice, directly from your Elgato Stream Deck.

FiveM Voice Controls is a push-to-talk voice recognition plugin that lets you execute FiveM client-console commands without needing to type them manually.

Simply hold a Stream Deck key, speak your command, and release. Your speech is recognised locally on your PC and sent directly to the FiveM client console.

The Full version includes its own bundled speech recognition engine. No Python installation or additional speech recognition setup is required.

Features

Push-to-talk: Hold a Stream Deck key to record and release to process your command.

Local speech recognition: Audio is processed on your PC, not sent to a cloud recognition service.

Custom voice commands: Configure your own spoken phrases and corresponding console commands.

NATO phonetic alphabet: Recognises phonetic letters for callsigns and vehicle registrations.

Spoken numbers: Converts spoken numbers into digits where appropriate.

Flexible arguments: Supports callsigns, registrations, vehicle descriptions, digits, and text.

Command validation: Configure required arguments, length limits, and cooldowns.

Confidence filtering: Reject transcriptions below your configured confidence threshold.

Microphone selection: Choose which microphone the plugin uses.

Status indicators: See when the plugin is listening, processing, sending, or encountering an error.

Example commands

Spoken phrase | Command sent

sam Whiskey November Tango One One | sam WNT11

go to 216 | p 216

p 216 | p 216

Check plate Alpha Bravo Five Charlie One | qvr AB5C1

10-3 Kilo Eight | 10-3 K8

10-7 Postal One Two Three | 10-7 123

3t red sedan plate Alpha Bravo Five Charlie One | 3t red sedan AB5C1

These are examples using the default command configuration. Your own commands and trigger phrases can be configured through the action settings.

Getting started

Install the plugin through Stream Deck.

Add Voice Command to a Stream Deck key.

Configure your microphone and commands.

Hold the key, speak, and release.

First launch: An internet connection is required to download the speech recognition model (approximately 150 MB). This is a one-time initial setup.

Requirements

Windows 10 or newer.

Elgato Stream Deck software 7.1 or newer.

A working microphone.

FiveM with its client-console connection available.

No Python installation is required for the Full version.

Documentation & Support

For detailed configuration instructions, troubleshooting, and advanced examples, visit:

GitHub – https://github.com/tigerjess12/fivem-voicecontrols

Enhanced Speech Recognition & Customisation

Implemented transcription confidence scoring in listener.py, allowing commands below a configurable confidence threshold to be rejected automatically.

Improved speech normalization with separate processing stages for corrections, spoken numbers, and NATO phonetic alphabet conversion.

Added detailed parsing traces to help identify how individual words are interpreted.

Expanded Voice Command Options

Added required argument settings and configurable minimum and maximum argument lengths.

Improved command matching to prioritise the longest matching trigger phrase, allowing more complex phrases and aliases.

Improved callsign and registration validation to reject invalid or unreadable input.

Updated the default QVR command to use plate recognition rather than callsign formatting.