Hermes Agent Browser Use is a real, first-party integration: Browser Use is now the default browser driver inside Hermes Agent, and Browser Use Cloud is one of the cloud browsers Hermes can rent for you.
You don't need a plugin, and you don't need an MCP server to get it working.
I expected to find a community add-on when I started checking this.
What I found instead is that Nous Research built it straight into the agent.
In this walkthrough I'll show you what is actually wired up, the five steps to switch it on, three browser tasks you can copy, and the safety rules I'd follow before I let any agent near a logged-in tab.
Everything below comes from the official Hermes docs, Browser Use's own docs, and read-only commands I ran on my own install on 12 October 2026.
I haven't published a video on Browser Use mode yet, so the video below is my one-click Hermes setup, which is the starting point for everything here.
What Hermes Agent Browser Use Actually Means
People type this keyword for two different reasons.
Some of you want to know whether Hermes Agent works with Browser Use, the open-source browser automation project at github.com/browser-use/browser-use.
The rest of you want to know how to get Hermes to use a browser for real jobs.
This post answers both, in that order.
Here is the short version of the first answer.
| Question | What I verified on 12 October 2026 |
|---|---|
| Is there an official integration? | Yes, the Hermes docs call Browser Use mode the default browser driver. |
| Is it a plugin? | No, a search of the curated plugin catalog for "browser-use" returned no entries. |
| Is it an MCP server? | No, Browser Use is not in the list that hermes mcp catalog prints. |
| Who made Browser Use? | The Browser Use team, and the main repo is MIT licensed. |
| Who made Hermes Agent? | Nous Research, and Hermes Agent is also MIT licensed. |
| Does it cost money? | Local browsing is free, and cloud browsers are pay as you go. |
Browser Use's own product page says the same thing from the other side.
It states that Hermes Agent ships Browser Use as its default browser driver, and that the Nous Portal Tool Gateway routes Hermes cloud browsing to Browser Use Cloud.
So both companies describe the same integration, which is the kind of agreement I look for before I write a setup guide.
The Three Real Routes
There are three supported ways to pair the two tools, and they stack on top of each other.
| Route | What it is | What you need |
|---|---|---|
| Browser Use mode | It is the default driver, and it gives the model one tool called browser_exec. |
You need a current Hermes build and nothing else. |
| Browser Use Cloud | It is a rented cloud browser with stealth, proxies and CAPTCHA solving. | You need a BROWSER_USE_API_KEY. |
| Nous Subscription | It is the same Browser Use cloud browser, billed through your Nous Portal plan. | You need a paid Nous Portal subscription. |
The driver and the browser are two separate choices.
The driver is how Hermes controls a page.
The browser is where that page actually runs.
Browser Use mode is a driver, and it can steer Hermes's own local Chromium, a Browser Use cloud browser, Browserbase or Firecrawl.
I explain the wider capability and every backend in my Hermes Agent browser explainer, so I'll keep this page on Browser Use.
🔥 Want my exact Hermes Agent Browser Use 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
How Browser Use Mode Works Inside Hermes
The older Hermes browser toolset is a set of separate tools, such as browser_navigate, browser_click and browser_type.
Each click is one tool call, and each call costs you tokens.
Browser Use mode swaps that set for a single tool called browser_exec.
The docs say the agent writes and runs Python in the browser to click, type, drag and scrape.
That means one tool call can do a whole chain of steps.
The engine underneath is Browser Harness, and the Browser Use CLI is a thin wrapper around it.
The Hermes docs say it is a regular Python dependency, so every current install ships with it, including the desktop app.
There are three details worth knowing before you start.
- Local browsing uses the Chromium that Hermes packages, and it never touches your own Chrome.
- The local browser starts in a clean, throwaway profile, so the agent is logged into nothing by default.
- The
browser_exectool is only offered to sessions that also have terminal access, because it runs model-written Python on your machine.
That last point matters for bots.
A locked-down Telegram or Discord surface without the terminal toolset keeps the older browser tools instead.
My wider Hermes toolkit, including the computer use agents that sit next to the browser tools, is in this video.
Hermes Agent Browser Use Setup In 5 Steps
Step 1: Get Hermes Installed And Current
If Hermes isn't on your machine yet, follow my one-click Hermes Agent setup first.
Then update it.
hermes update
hermes --version
This step matters more than it looks.
My own install is a late-September build of v0.21.4, and that build still treats the Browser Use CLI as a separate download inside the tools menu.
The current docs say the engine now ships with every install.
So an update is the difference between "it just works" and a one-off install prompt.
Step 2: Pick Your Browser In The Tools Menu
Run the tools menu and open the browser section.
hermes tools
Choose Browser Automation.
On a fresh install the first row is Local Browser, which is free and needs no API key.
That is the row I'd start with.
The same menu also lists Lightpanda, the Nous Subscription cloud browser, Camofox and Browser Use, among others.
I rank those options against each other in my best browser for Hermes list.
Step 3: Confirm Browser Use Mode Is On
You don't have to set anything for the default.
When browser.backend is unset, Hermes uses Browser Use mode automatically.
If you want to force it, add this to your config file.
# ~/.hermes/config.yaml
browser:
backend: "browser-use"
If you want the older click-by-click tools back, type /browser use off in a chat, or set the backend to "off".
browser:
backend: "off"
Camofox is the one exception.
The docs say Camofox has no endpoint for the harness to attach to, so Camofox setups keep the older tools automatically.
Step 4: Add Browser Use Cloud (Optional)
You only need this step for sites that block normal automated browsers, or for a server with no screen.
Create an API key in the Browser Use Cloud dashboard.
Then run hermes tools, open Browser Automation, choose Browser Use and paste the key when it asks.
The manual version is two small edits.
# ~/.hermes/.env
BROWSER_USE_API_KEY=your_key_here
# ~/.hermes/config.yaml
browser:
cloud_provider: browser-use
The Hermes docs add one warning that saves a lot of confusion.
The key only supplies the credential, and the menu choice decides which browser is active.
Once you have picked a provider, adding or removing a key does not switch it.
If you already pay for Nous Portal, you can skip the key and choose the Nous Subscription row, which runs the same Browser Use cloud browser on your existing bill.
Step 5: Run A First Test
Start a chat with hermes and give it something harmless.
Open github.com/trending and list the top five repositories with one line on each.
You should see a browser tool call, then an answer with real repo names.
If you get a notice saying the Browser Use CLI was not found, Hermes has fallen back to the older tools.
Run hermes update, or run hermes tools and choose Browser Use under Browser Automation to install it.
You can also run hermes doctor to see what the browser section is missing.
🔥 Want my Hermes setups as I build them? Inside the AI Profit Boardroom, I share my agent walkthroughs step by step, and there are four coaching calls a week where you can bring your own terminal errors. Over 3,400 members are building real automations in there. → Get access here
Browser Use As An Activity: 3 Tasks To Copy
Now for the second meaning of Hermes Agent Browser Use, which is the agent using a browser to get work done.
These are prompts to start from, and they are not results I'm claiming.
Run each one on a low-risk site first.
Task 1: Research A Live Page
Open the pricing pages for these three tools, and build a table of plan names and monthly prices.
Give me the page address next to every number.
Ask for the address next to every fact, because it makes checking fast.
The Hermes docs give one honest tip here.
For simple reading, web_search and web_extract are faster and cheaper than a browser.
Save the browser for pages that need clicks, logins or scrolling.
Task 2: Fill A Form And Stop Before Submit
Go to the contact form on this page and fill it in with the details below.
Do not press submit.
Take a screenshot and show me what you entered.
That "do not press submit" line is your approval gate.
As far as I can tell from the docs, Hermes does not pause before every click, so you have to build the pause into the task.
Turn on a visible window while you are learning.
browser:
headed: true
Headed mode opens a real Chromium window and keeps it open between turns, so you can watch and step in.
Task 3: A Logged-In Check
This is the one to be careful with.
Hermes gives you two ways to browse as yourself.
- Real profile browsing copies your active Chrome profile into a managed snapshot and drives that copy.
- The
/browser connectcommand attaches Hermes to a Chrome window over the Chrome DevTools Protocol, and it only works in the terminal CLI.
Real profile browsing is one line of config.
browser:
use_real_profile: true
The docs call this a consent-gated convenience and not an isolation boundary.
A page the agent visits runs with your real logins.
Quit Chrome before you start the session, because a running browser can lock the login databases.
Then keep the task read-only.
Open my analytics dashboard and tell me yesterday's total visits.
Do not change any settings and do not click anything that saves, sends or deletes.
If a site asks for a password the agent doesn't have, Hermes shows a masked prompt.
The docs say the password is encrypted on your machine, tied to that one site, and typed into the page without the model ever seeing it.
Approvals And Safety For Hermes Agent Browser Use
Hermes has real guardrails, but you should know exactly what each one covers.
| Guardrail | What it does |
|---|---|
| Command approvals | The approvals.mode setting gates dangerous shell commands, with smart, manual and off modes. |
| Card fills | Every saved-card fill asks you first, and sessions with nobody present are refused. |
| Real profile | It is off by default, and switching it off deletes the copied profile. |
| Closing your browser | Hermes never closes your browser without asking you first. |
| Website blocklist | You can list sites the web and browser tools must never open. |
| Private addresses | Every web address is checked first, and a cloud browser never sees your local network pages by default. |
Here is my own rule set on top of that.
- Keep
approvals.modeon smart or manual, and never run browser jobs in YOLO mode. - Give the agent a clean profile unless a task truly needs your logins.
- Never let an agent complete a payment, and do that final click yourself.
- Don't ask it to break a CAPTCHA on a site whose terms forbid automation.
- Write "stop and show me" into any task that sends, submits, books or deletes.
If you'd like a second pair of eyes on your site and your automation plan, you can book a free SEO strategy session with my team.
What It Costs
| Item | Cost |
|---|---|
| Hermes Agent | It is free and open source. |
| Browser Use mode with a local browser | It is free. |
| Browser Use Cloud browser | Browser Use lists it at $0.02 per browser-hour, metered by the minute. |
| Managed residential proxy | Browser Use lists it at $5 per GB. |
| Nous Subscription browser | It is billed as usage against a paid Nous Portal plan. |
| The model | You pay your model provider, unless you use a free or local model. |
Browser Use says there are no subscriptions and that purchased credits don't expire.
Its README and its pricing page disagreed about the size of the free signup credit when I checked, so I'm leaving that number out.
Check the live pricing page before you budget anything.
Honest Limits
- The older browser tools read pages as text, so very visual pages can confuse them.
- Page snapshots are cut at 15,000 characters by default, and the full version is saved to a file.
- The docs list file downloads from the browser as unsupported for the built-in tools.
- The
/browser connectcommand does not work from Telegram, Discord or the web interface. - Lightpanda has no screenshots, because it has no graphical renderer.
- Weaker models struggle with Browser Use mode, because they have to write working Python.
That last limit comes from Nous Research's own benchmark notes in the Hermes repo.
Their test showed big token savings on strong models and no real saving on a small local coder.
I break those numbers down in the GoldieBench version of this post, which is linked below.
The MCP Route And Jev Ultrafast
Browser Use does publish MCP servers.
The hosted one lives at https://api.browser-use.com/v3/mcp and takes your key in an x-browser-use-api-key header.
The local one starts with uvx --from 'browser-use[cli]' browser-use --mcp and needs your own model keys.
Hermes can add custom MCP servers with hermes mcp add, which accepts a --url and a header auth option.
I have not tested that pairing, and Browser Use's MCP page lists Claude, Cursor and Windsurf but not Hermes.
The built-in route is simpler, so I'd only use MCP if you want Browser Use's hosted agent to run a whole task for you.
Browser Use's own Hermes page also shows a standalone CLI install, which is uv tool install browser-use followed by browser-use skill install.
One related project is worth a mention.
Jev Ultrafast is Browser Use's separate browser agent that uses TypeSafe's Jev model to pick every click.
It is not wired into Hermes, and I cover it in my Jev Ultrafast GitHub walkthrough.
🔥 Ready to run agents that actually finish the job? The AI Profit Boardroom has my Agent OS walkthroughs, the Hermes tutorials as they ship, and four coaching calls a week where you can bring your own setup. Over 3,400 members are building with it right now. → Join the AI Profit Boardroom
Related Reading
- Hermes Agent browser: what the capability is and how to enable it covers the toolset and every backend.
- Best browser for Hermes ranks the browsers and backends you can pair with it.
- Jev Ultrafast GitHub explains Browser Use's separate Jev-powered browser agent.
- Best Hermes Agent tools shows where browser automation sits in my wider tool list.
Also On Our Network
- 🌐 Seven browser jobs a business can hand to Hermes Agent
- 🌐 Every Hermes Browser Use feature, with my verdict and scores
- 🌐 The Browser Use quick start inside an Agent OS
- 🌐 Which benchmarked models handle browser work best
FAQ: Hermes Agent Browser Use
Does Hermes Agent work with Browser Use?
Yes, and it is built in.
The official Hermes docs describe Browser Use mode as the default browser driver, and they list Browser Use Cloud as a supported cloud browser provider.
You don't need a plugin or an MCP server for either one.
How do I turn on Browser Use in Hermes Agent?
Run hermes update so you are on a current build, then run hermes tools and open Browser Automation.
Pick Local Browser for free local browsing, or pick Browser Use and paste a BROWSER_USE_API_KEY for cloud browsers.
Browser Use mode is the default driver, so there is no extra switch on a current install.
Is there a Browser Use MCP server for Hermes?
Browser Use publishes a hosted MCP server and a local one, but neither is in the Hermes MCP catalog.
I ran hermes mcp catalog on 12 October 2026 and Browser Use was not listed.
You can add it as a custom server with hermes mcp add, but the built-in integration is the simpler route.
Is Hermes Agent Browser Use free?
Local browsing is free, because Hermes drives its own packaged Chromium on your machine.
Browser Use Cloud is pay as you go, and Browser Use lists cloud browser sessions at $0.02 per browser-hour plus network usage.
You always pay for the model that does the thinking unless you run a free or local one.
Can Hermes use my logged-in Chrome?
Yes, it can do that in two ways.
You can set browser.use_real_profile to true, which makes Hermes browse on a managed copy of your active Chrome profile.
You can also type /browser connect in the terminal CLI to attach Hermes to a Chrome window over the Chrome DevTools Protocol.
Can Hermes Agent solve CAPTCHAs with Browser Use?
Browser Use Cloud includes stealth browsers, residential proxies and CAPTCHA solving.
Browser Use's own README says results depend on the site and that no browser setup guarantees every CAPTCHA is solved.
On a local browser you should expect to solve the challenge yourself in headed mode.
📺 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 AI entrepreneur, SEO expert, and founder of the AI Profit Boardroom, which has 3,400+ members.
I help business owners scale with AI agents, automation, and SEO.
- I have 400,000+ YouTube subscribers who watch my AI tool tests every week.
- I built a 7-figure agency from the ground up.
- I run daily AI training inside the Boardroom.
- I wrote "Link Building Mastery", my book on link building strategy.
→ Get my best AI training inside the AI Profit Boardroom
So here is the whole thing in one breath.
Update Hermes, pick a browser in the tools menu, test one harmless task, and add your own stop points, and you'll have a safe working setup for Hermes Agent Browser Use.
