The Hermes agent robot people are searching for is the Autonomous Lamp, a desk robot from Autonomous that can run Hermes Agent as its brain through an open-source operating system called lampOS.
Autonomous calls it a "fully-assembled open source companion robot", and it has a camera, microphones, speakers, a light and a five-servo arm that lets it nod, tilt and turn to look at you.
On 24 October 2026, Autonomous's official Lamp page listed it at $999 for a single unit, with free shipping.
Hermes is one of six supported agents, alongside Claude Code, Codex, OpenClaw, OpenCode and PicoClaw, and the Lamp's own device file in the lampOS repo sets Hermes as the default.
I covered it in my video using Autonomous's own footage, but I don't own a Lamp and I haven't tested one, so every step below comes from Autonomous's page and the lampOS README.
This page is the operator walkthrough, which means what you need, the setup steps, the brain, the character, the skills and the safety limits, in that order.
The Hermes Agent Robot At A Glance
This table is the whole product on one screen, so you know what you're dealing with before you plug anything in.
| Question | The answer from Autonomous's page and the lampOS README |
|---|---|
| What is it? | It is the Autonomous Lamp, a robotic desk lamp that Autonomous sells as a fully-assembled open-source companion robot. |
| Who makes it? | Autonomous makes it, and its footer lists NYC, SF, Shenzhen and Saigon. |
| What does it cost? | The official page listed $999 for one unit with free shipping on 24 October 2026. |
| What runs on it? | It runs lampOS, which the repo describes as the open-source operating system for physical AI. |
| Which agents can drive it? | The README lists Hermes, Claude Code, Codex, OpenClaw, OpenCode and PicoClaw. |
| Which agent is the default? | The Lamp's ROBOT.md file sets Hermes as the default gateway. |
| How do you set it up? | You use the Autonomous app on iOS or Android and your Wi-Fi network. |
| How does it stay safe? | A SAFETY.md file sets limits that the OS enforces without asking the model. |
The repo is public at github.com/autonomous-ai/lampOS, and it had about 490 stars on 24 October 2026.
What The Hermes Agent Robot Actually Is
The simplest way to think about it is that Hermes gets a body.
Hermes Agent is Nous Research's open-source agent, and on most setups it lives in a terminal or a desktop app.
The Autonomous Lamp gives that same kind of agent eyes, ears, a voice and an arm it can move.
Autonomous's announcement on X, which I showed in my video, put it like this: "We just gave Hermes a body, one eye, and two ears."
The README calls the software "Physical AI Operating System" and says Autonomous is building the "Android" for physical AI agents.
That's the key idea for operators, because the OS sits between the agent and the hardware.
The agent never drives a servo directly.
Instead, a skill tells the agent what it can do, the agent replies with a hardware marker, and the OS checks that marker against the safety rules before anything moves.
The same OS also runs on Autonomous Intern, which is an always-on desk agent, and on Hugging Face's Reachy Mini.
That means a skill you write for the Lamp can install on any compatible robot that has the right parts.
🔥 Want my exact Hermes Agent Robot Setup setup, step by step? Inside the AI Profit Boardroom you get step-by-step video trainings, my installable Agent OS and four coaching calls a week with 3,400+ members building real automations. → Join the AI Profit Boardroom
What You Need Before You Start
Autonomous's FAQ answers this in one line, which is that you need Wi-Fi and the Autonomous app.
The app is on iOS and Android, and both links come straight from the README.
The FAQ says that once you sign in and connect the Lamp, it runs on Autonomous's hosted AI right away, including the brain, the voice, face recognition and mood.
It also says that if you'd rather bring your own agent or keys, you can switch after setup.
So you don't need a Hermes install on day one, and that surprised me when I read it.
You can get it talking first and swap the brain to Hermes once it's working.
The Lamp's Hardware, From The Official Spec Sheet
Here's what Autonomous lists on the Lamp page, so you know what your agent can actually use.
| Part | What the official page lists |
|---|---|
| Form | It is a robotic desk lamp. |
| Movement | It uses 5 position-feedback servos. |
| Vision | It has a wide-angle camera with face tracking and motion sensing. |
| Audio | It has a built-in microphone array and built-in stereo speakers. |
| Compute | It runs on an 8-core ARM64 CPU. |
| Memory | It has 6 GB of RAM. |
| Connectivity | It connects over Wi-Fi 6. |
| Power | It takes 12V 5A DC. |
| Size | It measures 7.87 × 7.87 × 18.43 inches. |
| Weight | It weighs 6.6 lbs. |
| Warranty | It comes with a 1-year warranty. |
The page also lists room sensors for CO₂, airborne particles from PM1.0 to PM10, temperature and humidity.
One line on the page calls these "Lamp Pro's room sensors", so check which model you're ordering if room sensing matters to you.
Step 1: Add The Lamp In The Autonomous App
This is the first step in the README's quick start.
Open the Autonomous app and tap Add robot, then Lamp.
That's the whole step, and the app takes you straight into the next one.
Step 2: Set Up Wi-Fi
You pick your home or office network in the app.
The README says the app joins the robot's hotspot and hands over the keys and the pairing for you.
If your office network blocks new devices, sort that out with whoever runs it before the Lamp arrives.
Step 3: Talk To It
The README describes the first interaction simply.
You say something, the Lamp turns to look at you, the ring lights up and it answers.
The Lamp's device file turns on a wake word out of the box, because a lamp in a shared room hears every conversation in it.
That file lists "hey lamp", "hey autonomous" and "hey" plus your agent's name as the phrases it waits for.
Autonomous's FAQ adds that you can use a supported wake phrase or the attention control set up on your Lamp.
The README also gives a short list of things to try on a configured Lamp.
| Say or type | What the README says should happen |
|---|---|
| "Turn off the lights." | Lamp turns off its light through a local command. |
| "This lamp is too bright." | Lamp cuts the current brightness by half. |
| "You're speaking too loudly." | Lamp cuts the speaker volume by half. |
| "Make this lamp violet." | Lamp sets a solid purple light. |
| "I need light to read a book." | Lamp switches on its reading light scene. |
Those simple commands are handled by local rules and an intent fallback before they ever reach the main agent.
Anything the OS isn't confident about goes to the main agent, which is where Hermes comes in.
Step 4: Install A Skill From The Skill Store
The README says you install a skill from the Skill Store in one tap, and it's live on the next conversation.
A skill in lampOS is one folder with one SKILL.md file inside, written in markdown the agent reads.
Skills that need hardware declare the capabilities they require, so they only install on robots that have the right parts.
My advice is to install one existing skill, try it, and only then start changing things.
🔥 Want my Hermes setup done with you, step by step? Inside the AI Profit Boardroom, I've got a full Hermes Agent course plus the Agent OS, with video walkthroughs on profiles, skills and memory. You'll be learning alongside 3,400+ members with four coaching calls every week. → Get access here
Step 5: Build Your Own Skill
The README says you type what you want the Lamp to do in the app, and it writes the skill for you.
That's the no-code route, and it's where most people should start.
If you'd rather write one by hand, the README points to a skill creator and an example tracking skill in the repo.
Hardware skills emit markers in the form [HW:/path:{json}] instead of touching a servo bus or a GPIO pin.
The OS strips those markers out of the reply and sends them to the hardware layer before the words are spoken.
That's why the agent can say something and move at the same time without you writing any motor code.
Autonomous's FAQ is honest that custom behaviours, app connections and experiences for other people may need code and testing.
Step 6: Give It A Character With SOUL.md
The last step in the README's quick start is to edit SOUL.md.
The README says that once you edit it, the Lamp "is someone else on the next turn".
SOUL.md is the self of the robot, which covers who it is and how it talks.
Autonomous's page adds the fun parts, which are a name, a voice from a supported speech service and a sense of humour.
It also suggests a signature, such as a greeting, a recorded gesture or a little light.
How To Make Hermes The Brain
The README's agent section says you run Hermes, Claude Code, Codex, OpenClaw, OpenCode or PicoClaw, then choose a compatible model and configure its voice, personality and tools.
All six sit behind one gateway in the runtime layer, and the README says you can switch between them live from the web UI.
It also says the persona, memory and connectors move with it when you switch.
The Lamp's device file names Hermes as the default gateway, so Hermes is the natural first brain to try.
The OS itself runs on the robot, but the README is clear that model calls may use remote services, depending on how you configure it.
| Where the brain runs | What Autonomous says |
|---|---|
| Autonomous's hosted AI | The FAQ says the Lamp runs on hosted AI right after setup. |
| A cloud model you choose | The page says you can connect a cloud model from a provider you pick. |
| A local model | The FAQ says yes with a compatible local model endpoint. |
| Fully offline | The FAQ says you also need compatible speech and vision services on your own hardware. |
If you want the Hermes side of this in more depth, my guide to Hermes Agent v0.22 covers the current release.
How Lamp Hands Work To Your Computer
This is the part that got my attention as an operator.
The README describes a skill called harness-use that sends coding and research tasks to agents already running in Autonomous's Harness app on a paired computer.
You ask a named agent to work, answer its follow-up questions, and get the result back through voice or chat.
Pairing happens from OS Monitor and Harness Desktop on the same local network.
There's also a computer-use skill that lets the Lamp's agent inspect a Mac app, click or type, and check the result through a paired Autonomous Buddy.
The README says Buddy bundles Cua Driver for those observations and actions.
If computer-use agents are new to you, my post on Hermes CUA explains how that world works.
The Safety Limits, And Where They Live
Safety in lampOS lives in a file called SAFETY.md, and the README calls the safety gate "a pure function" of that file.
It sits below the agent and in every request path, with no model in the loop.
That means the same limit applies whoever asked, whether that's Hermes, a skill or the web UI.
Here's what the Lamp's own SAFETY.md in the repo sets on 24 October 2026.
| Limit | What the Lamp's SAFETY.md says |
|---|---|
| Brightness | The daytime ceiling is 120 on a 0 to 255 scale, and higher requests are clamped. |
| Quiet hours | From 22:00 to 07:00, the light ring dims to a ceiling of 40. |
| Movement speed | Joints are capped at 120 degrees per second, so fast moves are stretched out. |
| Stop | Stop and release are always available and never wait behind the agent. |
| Volume | The volume ceiling is set to 100 percent in the shipped file. |
The README links to a separate safety doc that lists what the gate does not cover yet, and I'd read it before putting a Lamp near customers or kids.
If you want to talk through where an agent with a body could fit your business, you can book a free SEO strategy session with my team.
Hermes Agent Robot Limits To Plan Around
These are the edges I'd want to know about before ordering one.
| Limit | What I'd do about it |
|---|---|
| The default setup uses Autonomous's hosted AI. | I'd decide early whether I'm happy with that or want my own agent and keys. |
| A local model alone doesn't make the Lamp offline. | I'd plan for local speech and vision services if offline matters to me. |
| Custom behaviours may need code and testing. | I'd start with Skill Store skills and the app's skill writer first. |
| Room sensors are described as a Lamp Pro feature in one line of the page. | I'd confirm the model before ordering for air-quality jobs. |
| Face or voice recognition needs enrolment and compatible services. | I'd check what my chosen setup supports before promising it to anyone. |
| The hardware layer is GPL-3.0 while the rest is Apache-2.0. | I'd read the licence section before shipping a commercial driver. |
The README's licence section says everything outside hal/ is Apache-2.0 and hal/ is GPL-3.0.
My Take On The Hermes Agent Robot
The big shift here is that your agent is no longer stuck behind a screen.
It can see who walked in, answer out loud, nod, change the light and hand a coding task to your computer.
The part I like most as an operator is that safety lives in a file the OS enforces, not in a prompt the model might ignore.
The part I'd be careful with is expectations, because I haven't tested one and the setup depends on which brain and services you choose.
🔥 Want help wiring Hermes into your real work? The AI Profit Boardroom has Hermes Agent training, the Agent OS and four weekly coaching calls where you can ask me which skills to build first. → Join the AI Profit Boardroom
Related Reading
- The Hermes computer use skill, set up step by step
- Hermes CUA, the plugin that gives Hermes its own throwaway computer
- The best Hermes Agent use cases right now
- Hermes Meshy, the plugin that lets Hermes make 3D models
Also On Our Network
- 🌐 Seven real business uses for the Hermes agent robot
- 🌐 The Autonomous Lamp reviewed and scored feature by feature
- 🌐 The Hermes agent robot quick start for an Agent OS
- 🌐 The models behind the Hermes agent robot, measured on GoldieBench
FAQ: Hermes Agent Robot Setup
What is the Hermes agent robot?
It's the Autonomous Lamp, an open-source desk robot from Autonomous that can run Hermes Agent as its brain through lampOS.
How much does the Autonomous Lamp cost?
Autonomous's official page listed it at $999 for one unit with free shipping on 24 October 2026.
How do I set up the Autonomous Lamp?
You add it in the Autonomous app, set up Wi-Fi, talk to it, install a skill, build your own skill and edit SOUL.md.
Can I run Hermes on the Lamp with a local model?
Autonomous says yes with a compatible local model endpoint, but fully local conversation also needs local speech and vision services.
Do I need to code to use the Hermes agent robot?
You don't need to code to start, because voice, personality and Skill Store skills work without it.
How does the Lamp stay safe?
The OS enforces SAFETY.md below the agent, which caps brightness, sets quiet hours, limits speed and keeps stop always available.
If you'd like my team to look at how agents like this could fit your business, book a free SEO strategy session and we'll talk it through.
📺 Video notes + links to the tools 👉
🎥 Learn how I make these videos 👉
🆓 Get a FREE AI Course + Community + 1,000 AI Agents 👉
About Julian
I'm Julian Goldie, an SEO entrepreneur, author and founder of the AI Profit Boardroom, which has 3,400+ members.
I help business owners scale with AI agents, automation and SEO.
- I've built a 7-figure SEO and link-building agency, Goldie Agency, with 70+ people in Bangkok.
- I've grown a YouTube channel to 400,000+ subscribers.
- I wrote the book "Link Building Mastery", which covers nine link building strategies.
- I run GoldieBench, where I score AI models on real one-shot builds.
→ Get my best AI training inside the AI Profit Boardroom
Add it in the app, connect Wi-Fi, install one skill, give it a soul and check the safety file, and you'll be up and running with the Hermes agent robot.
