Hermes Voice Mode

Overview

Voice Mode is a built-in feature shipped with Hermes Agent v0.20 ("The Herald Release," August 3, 2026). It enables real-time conversational voice interaction with your Hermes agent — the agent speaks while generating responses and can be interrupted mid-sentence. Combined with wake word support, this makes Hermes a hands-free assistant across the CLI, TUI, and desktop app.

What it does

• Real-time voice conversation: Speak to Hermes naturally; it responds with synthesized speech while generating, not after
• Mid-sentence interruption: Cut the agent off if it is going in the wrong direction — no need to wait for it to finish
• Wake word activation: Say "hey hermes" (or any configured phrase) to start a fresh voice session without touching the keyboard
• TTS output modes: Choose when the agent speaks — only when you send voice, or for every message
• Cross-platform: Works in the CLI, TUI, and desktop app with consistent behavior

Modes

Mode	Command	Behavior
off	/voice off	Text only (default)
voice_only	/voice on	Speaks reply only when you send a voice message
all	/voice tts	Speaks reply to every message

Voice mode setting is persisted across gateway restarts.

Quick start

1. Update to Hermes Agent v0.20 or later
2. Ensure a microphone is connected and a TTS provider is configured
3. Toggle voice mode: /voice on for voice-only responses, or /voice tts for all responses
4. Set a wake word in your configuration to enable hands-free session start
5. Speak naturally — interrupt by speaking while the agent is talking

Use cases

• Hands-free coding: Dictate changes, ask questions, and get spoken answers while your hands are on the keyboard doing other work
• Accessibility: Voice interaction for users who cannot or prefer not to type
• Mobile or remote work: Interact with your agent from across the room or while walking
• Pair programming: Have Hermes narrate its analysis while you review code
• Meeting workflows: Ask Hermes questions during a call without switching to the keyboard

Limitations

• Microphone required: Voice input needs a working microphone; quality affects transcription accuracy
• TTS provider dependency: Natural-sounding speech requires a configured TTS provider — system TTS works but is less natural
• Latency: Real-time voice adds streaming latency; not suitable for latency-critical workflows
• Noisy environments: Background noise degrades wake word reliability and transcription quality
• Privacy: Voice data is sent to speech-to-text providers — review your provider's data policies

Related

• Hermes Agent Documentation: Voice Mode
• Talk to Hermes — pairs well with voice for hands-free research