Hermes Agent Voice Mode Complete Guide 2026: Hands-Free CLI, Telegram Voice Replies, and Discord Voice Chat

---

"What if you could talk to your AI agent without touching the keyboard?"

With Hermes Agent's voice mode, that becomes reality. Hands-free operation in the CLI, voice replies on Telegram, and live conversation in Discord voice channels — three ways to use it, all in one agent.

In this article, based on the Hermes Agent official Voice Mode guide, we explain "what you can do" and "how to set it up" with concrete steps, beginner-friendly.

Here's what you'll learn:

• The 3 ways to use voice mode and which to choose
• Required packages, system dependencies, and choosing STT/TTS providers
• Concrete setup steps for macOS/Ubuntu
• The free "local STT + Edge TTS" setup

---

What You Can Do with Voice Mode (3 Experiences)

The official docs classify voice mode into "three different experiences."

The official recommended staged rollout:

1. Get text working perfectly first
2. Enable voice replies
3. Move to Discord voice channels (if you want the full experience)

---

What's Convenient (Concrete Use Cases)

• Hands-free CLI: verbally instruct "debug this" or "read this file" without touching the keyboard
• Ideas while walking: ask by voice during a walk, get answers by voice
• Telegram voice replies: send a voice message from your phone's Telegram, and Hermes replies by voice
• Discord voice channels: keep Hermes in your team VC and let it join the conversation

Important: with Nous Portal, a single OAuth covers both LLM and TTS. No extra API key setup needed.

---

Setup Steps (5 Steps)

Step 1: Confirm text mode works

Before voice, solidify the basics.

hermes doctor    # verify Hermes starts
hermes chat      # verify you can converse normally in text

If it answers a simple question like "Hello, what can you do?" in text, you're good.

Step 2: Install the extras you need

Packages differ depending on your use case.

# CLI mic + audio playback (hands-free operation)
pip install "hermes-agent[audio]"

# Messaging platforms (Telegram/Discord)
pip install "hermes-agent[messaging]"

# High-quality TTS (ElevenLabs)
pip install "hermes-agent[elevenlabs]"

# Local NeuTTS (optional, fully local)
pip install "hermes-agent[neutts]"

# Everything
pip install "hermes-agent[all]"

Step 3: Install system dependencies

macOS

brew install portaudio ffmpeg opus espeak-ng

Ubuntu / Debian

sudo apt install portaudio19-dev ffmpeg libopus-dev espeak-ng

Role of each package:

• portaudio → mic input and audio playback (for CLI voice mode)
• ffmpeg → TTS audio conversion and messaging delivery
• opus → Discord voice codec
• espeak-ng → phoneme conversion backend for NeuTTS

Step 4: Choose STT (speech-to-text) and TTS (text-to-speech) providers

Hermes supports both local and cloud.

Easiest and cheapest setup (recommended for beginners):

• STT provider: whisper (local)
• TTS provider: edge (free)

Add to .env:

HERMES_STT_PROVIDER=whisper
HERMES_TTS_PROVIDER=edge

Provider selection guide:

Step 5: Launch and talk

# CLI mic mode (interactive voice loop)
hermes voice

# Normal start (voice replies enabled on Telegram/Discord)
hermes start

In CLI mode, speak and it replies by voice instantly. End the conversation with Ctrl+C.

---

Discord Voice Channel Setup (Advanced)

Using Discord VC requires separate configuration. See the official docs. Overview:

• Grant the Discord Bot "connect to Voice Channel" permission
• opus package installation required
• Specify the voice channel ID in Hermes' Discord config

---

Ready-to-Use Checklist

• [ ] hermes doctor confirms normal startup
• [ ] You can converse normally in text chat
• [ ] portaudio / ffmpeg / opus / espeak-ng installed
• [ ] STT_PROVIDER / TTS_PROVIDER set in .env
• [ ] hermes voice runs in CLI mic mode

---

FAQ

Q1. Can I use voice mode for free?

Yes. With STT=whisper (local) + TTS=edge (free), it's completely free.

Q2. Does it support Japanese speech recognition?

whisper supports Japanese, and edge TTS provides natural Japanese read-aloud.

Q3. Can I talk by voice from my phone?

With Telegram integration, send a voice message in the Telegram app and Hermes processes it and can reply by voice.

Q4. Can it live in a Discord voice channel?

Yes. With dedicated settings, Hermes can reside in a Discord VC for live conversation.

Q5. Can I have multi-turn conversation in CLI voice mode?

Yes. hermes voice enters a dialogue loop — keep talking.

Q6. What about privacy?

With the whisper + neutts setup, audio processing is fully local. Cloud TTS sends data via API.

Q7. Which OS does it work on?

macOS and Linux. Windows via WSL.

Q8. Do I need Nous Portal?

Not required, but it covers LLM + TTS with one OAuth, making setup dramatically easier.

---

Summary — Voice Mode Makes Hermes Your "Partner"

Conclusion: Hermes Agent's voice mode is a practical, flexible feature that realizes "AI operation without typing."

• The free setup (whisper + edge) lets you start right now
• Expandable in stages: CLI → Telegram voice replies → Discord VC
• Japanese supported (both STT and TTS)

What you can do in 3 minutes:

pip install "hermes-agent[audio]"
brew install portaudio ffmpeg    # macOS
hermes voice                      # start talking

Asking "what's causing this bug?" while on a walk and getting the answer by voice — that's everyday life with Hermes voice mode.

---

Information in this article is as of July 2026. Content based on Hermes Agent official docs — Use Voice Mode with Hermes.