Hermes Agent ElevenLabs: Make Sound Effects With the Official Plugin

Share this post

Hermes Agent ElevenLabs support just got a new trick: Nous Research released an official plugin called elevenlabs-sound-effects that lets Hermes turn a short text description into a real sound effect. You type “a heavy wooden door creaking open in a stone hall” and Hermes makes the audio file and sends it to you in chat. The plugin (v1.2.0) was added to the Hermes plugin catalog on 26 October 2026, and Nous announced it on X the next day as part of a growing set of “Game Dev Studio” tools for Hermes. Below I cover what it does, how to install it, every setting, what it costs to run, and how it fits next to the ElevenLabs voice you may already use in Hermes. I’ve also put one of my Hermes walkthrough videos below.

Key takeaways

  • One tool, elevenlabs_sound_effect: text in, sound file out (0.5 to 30 seconds, optional seamless loop).
  • Install with hermes plugins install elevenlabs-sound-effects.
  • Uses the same ELEVENLABS_API_KEY as Hermes’ ElevenLabs voice.
  • Ships a sound-design skill, a free local ffmpeg script and a Desktop sound library.
  • Want this working in your business? Claim a free strategy session.

What the Hermes Agent ElevenLabs plugin does

The plugin connects Hermes to the ElevenLabs Sound Effects API. According to the catalog page, it adds exactly one tool to your agent: elevenlabs_sound_effect. You describe a sound in plain English and Hermes generates an audio file between 0.5 and 30 seconds long. It can also make the sound a seamless loop, which matters for things like background music, engine hums or rain.

Examples straight from the README:

  • “Glass shattering on a tile floor”
  • “A short retro coin pickup blip”
  • “90s hip-hop drum loop, 90 BPM, loopable”
  • “The sound of a heavy wooden door creaking open in a stone hall”

Hermes picks the details for each sound itself: the description (up to 450 characters — longer prompts are refused before anything is sent), the length, whether it loops, and a prompt_influence value from 0 to 1. A higher value follows your words more literally; a lower one gives more variety. You can also ask for a specific model on a single sound.

It’s an Official plugin, maintained by NousResearch, listed under Tools, and it needs Hermes 0.22.0 or newer. It runs on Linux, macOS and Windows, and its only Python dependency is httpx, which Hermes already ships.

How to install the ElevenLabs plugin in Hermes

Installing takes one command. Here are the steps, in order:

  • Make sure you’re on Hermes 0.22.0 or later (hermes update if you’re not sure).
  • Run hermes plugins install elevenlabs-sound-effects. You can also install it straight from the hermes-official-plugins repo with hermes plugins install NousResearch/hermes-official-plugins#elevenlabs-sound-effects.
  • If ELEVENLABS_API_KEY isn’t set yet, the installer asks for it. You create a key in ElevenLabs under Profile → API keys.
  • The tool only shows up for the agent once the key is set.

Two details catch people out:

  • Plugins are per profile. If you run several Hermes profiles, install it on each one you want, for example hermes -p <profile> plugins install elevenlabs-sound-effects.
  • Restricted keys need the right permission. If your ElevenLabs key is restricted, turn on the Sound Effects permission. It’s separate from the Text to Speech permission.

Prefer not to touch the terminal? Ask your agent: “Install the elevenlabs-sound-effects plugin on this profile and check the key works.”

Want Hermes making sounds, videos and content for your business? Inside the AI Profit Boardroom you get the Agent OS — Hermes, OpenClaw, OpenCode, Claude and Codex in one place — plus step-by-step Hermes training and four live coaching calls a week.

Want a plan for your business first? Claim your free strategy session.

Join the AI Profit Boardroom →Claim Your Free SEO Strategy Session →

Hermes ElevenLabs settings explained

In Hermes Desktop, open Capabilities → Plugins and click the gear icon on the plugin. Everywhere else, use hermes config set. These are the four settings the README lists:

SettingDefaultOptions and notes
api_key(unset)Stored in .env as ELEVENLABS_API_KEY
sound_modeleleven_text_to_sound_v2eleven_text_to_sound_v2 or eleven_text_to_sound_v3
output_formatmp3_44100_128mp3_44100_128, mp3_44100_192, opus_48000_128, opus_48000_192, wav_44100, wav_48000
api_base_urlhttps://api.elevenlabs.ioChange only for an ElevenLabs data-residency endpoint, e.g. https://api.eu.residency.elevenlabs.io

For example, to switch to the v3 sound model:

hermes config set plugins.entries.elevenlabs-sound-effects.settings.sound_model eleven_text_to_sound_v3

And to get 48 kHz WAV files:

hermes config set plugins.entries.elevenlabs-sound-effects.settings.output_format wav_48000

Watch your ElevenLabs plan. The README notes that ElevenLabs documents 192 kbps MP3 for the Creator plan and above, and 44.1 kHz WAV for Pro and above. If your plan refuses a format, the tool simply reports ElevenLabs’ error back to you.

Where your sounds go and how they reach you

Every sound is saved under your Hermes home in cache/generated/audio/, with a file name built from the description, a timestamp and an ID, so nothing is ever overwritten. Each generation also adds one line to a sounds.jsonl log in the plugin’s own data folder: the prompt, length, model, format, file size and the credits ElevenLabs billed. That log stays on your machine.

Delivery depends on where you’re chatting:

  • Telegram, Discord, Slack and other messaging apps: the tool hands the agent a MEDIA: line, and the file arrives as an attachment.
  • MP3 (the default) plays inline on most chat apps. Ogg Opus and WAV can arrive as plain files — Telegram shows them that way.
  • Discord: Hermes releases after v0.22.5 send MP3 and WAV as normal attachments that play inline.
  • Terminal: you get the file path.
  • Hermes Desktop: you get a sound card (below).

So if you run Hermes on a VPS and talk to it on Telegram, you can ask for a sound from your phone and get it back in the same chat.

The sound-design skill and the free ffmpeg script

The plugin also bundles a skill called elevenlabs-sound-design. It teaches Hermes how to write good ElevenLabs prompts (based on ElevenLabs’ own prompting guide), make loops, choose a format for where the sound is going — a game engine, a video, an app interface or a chat — and build a sound pack.

Inside that skill is sfx_tool.py, a local post-processing script that uses ffmpeg. It can downmix to mono, trim silence, normalise loudness, add fades, check a loop seam for clicks, convert between OGG, WAV and MP3, and pack a named set into a folder with a manifest.json and a zip. It can also build a game “variation set” with matched loudness, such as three gravel footsteps.

The important part: this processing costs no ElevenLabs credits, because it runs on your own machine. You pay ElevenLabs once to generate a sound, then tidy it up for free. One example from the README:

python3 sfx_tool.py pack t1.wav t2.wav t3.wav --name "Footstep Gravel" --out-dir . --variations

You don’t need to run this yourself. Just ask Hermes: “Make three gravel footstep variations, mono, levelled, packed as OGG.”

Want this set up for you instead of doing it yourself? Claim your free strategy session →

Claim Your Free SEO Strategy Session →

Hermes Desktop: sound card and sound library

If you use Hermes Desktop, the plugin adds two things. The Desktop half is opt-in: open Capabilities → Plugins, switch to list view and turn on Desktop: ElevenLabs Sound Effects. Without it, Desktop still plays every sound in its built-in player.

  • Sound card: each sound shows in chat with a waveform, play/pause, a gapless loop toggle so you can hear the seam, download in each format, reveal in your file manager, and “Next” actions — 3 variations, shorter, longer, make it loop, or try the other model. Each Next row shows its cost, and nothing is billed until the agent actually runs the tool.
  • Sound library: press ⌘K and choose ElevenLabs: open sound library. You get every sound the active profile made, with its prompt and settings, a format filter and a prompt search. You can replay, download, reuse a prompt or jump back to the chat that made it.

The sound card needs a Hermes Desktop with tool-media renderers; the README says that arrives in the first release after v0.22.5. Older Desktops use the built-in player instead. Running Hermes on a remote machine? Install the plugin there and also into your Desktop via Capabilities → Plugins → Install from Git, with Desktop UI checked.

What it costs and what leaves your machine

Every call is a paid ElevenLabs generation on your own ElevenLabs account. The README reports what it was billed while testing: about 10 credits per second when a length is set, and a flat amount (about 50 credits) per generation when ElevenLabs picks the length. Loops can come back a little longer than requested, and the tool reports the real length next to the one you asked for. Check ElevenLabs’ own pricing for your plan.

Privacy is simple: only the sound description and its parameters are sent to ElevenLabs, with your API key. Nothing else from your conversation goes with it. Format conversions and the processing script run locally.

Who should use the Hermes ElevenLabs sound plugin

Nous framed this release as part of a “Game Dev Studio” suite, so game builders are the obvious fit. But it’s useful well beyond games:

You make…Use it for
GamesFootsteps, pickups, UI blips, ambient loops and variation packs
Videos and ShortsWhooshes, hits, transitions and background beds
Apps and websitesShort interface sounds in the right format
Podcasts and adsIntros, stingers and scene sounds

If you already use Hermes for 3D work, it pairs naturally with the Meshy plugin I covered in Hermes Meshy: models from one plugin, sounds from the other, all from the same agent.

ElevenLabs sound effects vs ElevenLabs voice in Hermes

Hermes already supports ElevenLabs for text-to-speech — that’s the voice your agent speaks with. This plugin is different: it makes sound effects and music loops, not speech. The good news is they share the same key. If your ElevenLabs voice is already set up in Hermes, the installer finds ELEVENLABS_API_KEY and there’s nothing to enter.

For more on giving your agent a voice, read my guide to Hermes Apollo Voice.

My verdict on Hermes Agent ElevenLabs sound effects

This is a tidy, well-built plugin. One tool, clear settings, a skill that knows how to prompt, and a free local script for the fiddly editing work. The Desktop sound library is a nice touch for anyone making lots of sounds. Watch your credits, start with MP3, and if you build games, videos or apps, Hermes Agent ElevenLabs sound effects are worth installing today.

About Julian Goldie

I’m Julian Goldie — founder of Goldie Agency and the AI Profit Boardroom. I test new AI agents almost every day for 432K+ YouTube subscribers and run Hermes, OpenClaw, OpenCode, Claude and Codex side by side in my own Agent OS. Claim your free strategy session or join the AI Profit Boardroom.

Want another angle? Read the AI Profit Boardroom guide to Hermes Agent ElevenLabs.

FAQ: hermes agent elevenlabs

What is the Hermes Agent ElevenLabs plugin?

It’s the official elevenlabs-sound-effects plugin from Nous Research. It adds one tool, elevenlabs_sound_effect, that turns a text description into an audio file from 0.5 to 30 seconds, optionally as a seamless loop.

How do I install ElevenLabs sound effects in Hermes?

Run hermes plugins install elevenlabs-sound-effects. The installer asks for your ElevenLabs API key if it isn’t set. Plugins are per profile, so use hermes -p <profile> plugins install elevenlabs-sound-effects for other profiles.

Do I need a separate ElevenLabs key for sound effects?

No. It uses the same ELEVENLABS_API_KEY as Hermes’ ElevenLabs voice. If your key is restricted, enable the Sound Effects permission, which is separate from Text to Speech.

Does the Hermes ElevenLabs plugin cost money?

Each sound is a paid ElevenLabs generation billed to your ElevenLabs account. The bundled ffmpeg script and Desktop format conversions run locally and cost no credits.

Can Hermes send the sounds to Telegram or Discord?

Yes. The file is sent as an attachment on Telegram, Discord, Slack and other messaging platforms. MP3 plays inline on most chat apps; in Hermes Desktop it shows as a sound card.

Ready to put Hermes to work in your business? Join the AI Profit Boardroom for the Agent OS, setup training and daily updates.

Or claim your free strategy session and we’ll map out your setup.

Starting from zero? Get the free AI course and community.

Join the AI Profit Boardroom →Claim Your Free SEO Strategy Session →Get the FREE AI Course + Community →

Related reading

Last updated October 2026. This page is a living guide to hermes agent elevenlabs — the facts here move fast and it is updated as they do.

Done-for-you AI SEO

Rank #1 with AI SEO and get more leads, traffic and sales

Done-for-you SEO packages: backlinks, AI-powered articles and an account manager every month. Save 30% with code BLACKFRIDAY (ends Dec 1). No setup fees.

See SEO packages →or book a free strategy session

Table of contents

Related Articles

OpenClaw Enterprise explained: what the open-source agent control plane does, who it’s for, how it differs from OpenClaw, and how to start a pilot.
OpenCode Desktop plugins explained: what v2.2.0 adds, how GUI plugins work, the package setup, Vite build and install, plus the other plugin changes.
Hermes multiple profiles explained: run several Hermes agents on one machine with one gateway and desktop backend, each with its own keys, bots and jobs.
Hermes llama.cpp explained: Hermes Agent now installs llama.cpp, rates models for your GPU and runs them locally with no account, key or internet needed.