Hermes Workspace V2 dropped and I've been hammering it daily for weeks — this post is the proper walkthrough of what's new, what's different, and what you actually need to know before you touch it.
I'm not going to rehash what the workspace already is — I covered that in the original Hermes Workspace post and there's no point repeating myself.
This post is V2-specific.
What's Actually New In Hermes Workspace V2
Hermes Workspace V2 isn't a fresh paint job on the same thing.
It's a structural rebuild of how you talk to, manage, and observe Hermes agents from a single screen.
The biggest change is that chat is finally first-class — the original dashboard treated it as a side-feature, V2 builds the whole UI around it.
Memory is now editable inline instead of needing terminal commands.
Skills get a visual toggle interface where you can flip 2,000+ ready-made skills on and off without touching config files.
The inspector panel shows agent reasoning step-by-step, which used to require digging through log files.
Profiles let you run multiple agents side by side with separate memory and skill loadouts.
And the whole thing now runs as a progressive web app, so it's on your phone too.
If you're new to the broader workspace concept, start with Hermes Agent Workspace before this V2 deep-dive.
The V2 Release Notes In Plain English
Let me translate the official release notes into something useful.
The first headline change is the "native gateway" — V2 sits between you and Hermes as a proper interface layer, not just a UI shell that calls the CLI underneath.
That means responses are faster, the connection is more reliable, and the gateway can route requests to different model providers without you reconfiguring anything.
The second change is sub-agent orchestration baked into the UI.
In V1 you could technically run multiple agents, but you had to manage them through terminal commands and separate processes.
V2 has a profile system where each agent has its own session, skills, memory, and context — and you switch between them with a click.
The third change is the new visual workspace metaphor.
V2 ships with a "virtual office" view that shows your agents as visual entities working together.
You can swap between round-table view, grid view, and office view depending on what you're doing.
For me the round-table view is actually useful when I'm orchestrating a research swarm — I can see which agent is stuck and which one is producing output.
How V2 Changed The Install Flow
This is one of the biggest practical wins.
V1 had a finicky install where you needed to clone the repo, run npm install, manage Node versions, and hope nothing broke.
V2 has a "Hermes-assisted" install where you literally type "set up Hermes Workspace V2 for me" into your existing Hermes terminal and it does the rest.
That's not marketing fluff — it actually works.
The agent installs the dependencies, sets up the gateway, configures the connection to your existing Hermes setup, and tells you the URL to open.
If you'd rather do it manually, the Docker option is still there.
One command and you're running.
If you don't have Hermes installed yet, see How To Setup Hermes Agent first — V2 assumes you already have a working Hermes setup.
Want my full Hermes Workspace V2 setup playbook? Inside the AI Profit Boardroom, I've got step-by-step V2 walkthroughs, the exact profiles I use, and the skill configs that make it actually productive. Plus weekly coaching calls + 3,000+ members. Get access here for $59/mo (locked rate)
V2 Walkthrough — Screen By Screen
Let me walk you through what V2 actually looks like when you open it.
The top bar shows your active profile, the connected model provider, and a status indicator for the gateway connection.
The left sidebar is the navigation — Chat, Memory, Skills, Files, Terminal, Inspector, Tasks, and Settings.
The main panel changes based on what you've selected.
Chat Panel
This is where most of your time goes.
Full markdown rendering, code blocks with syntax highlighting, file uploads, and the agent can paste images and screenshots back at you when relevant.
You can scroll back through the full conversation history per profile, which V1 didn't preserve cleanly across sessions.
Memory Panel
Visual tree view of everything your agent remembers.
You can browse by folder, search across all entries, and edit memories inline.
You can also bulk-import memories from a markdown file or an Obsidian vault — which is a big deal if you've been building a Hermes second brain elsewhere.
Skills Panel
Card-based UI for the 2,000+ available skills.
Each skill has a toggle, a description, and an "edit prompt" button if you want to customise it.
You can create new skills inside the UI with a form — no more editing YAML files in a terminal.
Files Panel
Standard file browser, but with awareness of which files Hermes is currently using as context.
You can mark files as "always include" or "exclude" from agent context.
Terminal Panel
The CLI is still there for power users.
You can run any Hermes command and it shows up in the same UI.
Inspector Panel
This is the killer feature for debugging.
Live view of agent reasoning — every step, every tool call, every output, with timing and token counts.
If your agent does something weird, the inspector tells you exactly why.
Tasks Panel
Kanban-style board with Backlog, To Do, In Progress, Review, and Done columns.
Drag tasks between columns and agents work them in the background.
This is genuinely useful when you have one agent on research, one on writing, and one on quality control.
If you want a deeper dive on the task board specifically, Hermes Kanban covers it.
The Profile System (V2's Biggest Win)
The profile system is the V2 feature I use most.
In V1 you basically had one agent setup that you had to reconfigure every time you wanted to do something different.
In V2 you have profiles — each with its own session history, skill loadout, memory store, and context settings.
I have profiles for SEO research, content writing, code review, and personal admin.
Each one is essentially a different "personality" with its own knowledge and tools.
Switching between them takes one click.
This sounds small but it changes how you work with the agent — you stop fighting context bleed where the agent gets confused because last week's task is polluting this week's output.
What V2 Removed Or Changed From V1
Not every change is additive — V2 also retired some V1 patterns.
The old "global config" file is gone — settings now live per-profile.
The legacy dashboard at the V1 default port is deprecated, though it still works for now.
The plugin system from V1 has been folded into the new skills architecture, so old V1 plugins need to be ported to the V2 skill format.
This is the bit that catches people out — if you have custom V1 plugins, plan a migration day before you upgrade.
The good news is most plugins port in under an hour because the V2 skill schema is simpler.
V2 Compared To V1 (Honest Take)
Here's a clean side-by-side.
| Feature | V1 | V2 |
|---|---|---|
| Chat with agent | Limited | First-class |
| Memory editing | Terminal only | Visual editor |
| Skills toggle | Config files | UI toggles |
| Multi-agent | Manual | Profile system |
| Mobile access | No | Progressive web app |
| Inspector | Logs only | Live UI |
| Kanban board | No | Yes |
| Install | Manual npm | One-command |
| Gateway | None | Native gateway |
| Virtual office view | No | Yes |
For 95% of users, V2 is the obvious upgrade.
The 5% who should stay on V1 are people with deeply customised V1 setups that aren't worth porting.
For everyone else, the upgrade pays for itself in the first day.
How To Hear Other Users On V2
I did a live Q&A on V2 with the community — questions, complaints, real workflows, all of it.
The Q&A covers the upgrade path from V1, the migration gotchas, and how to use the profile system properly.
If you're on the fence, watch this before deciding.
Real Use Cases I'm Running In V2
These are actual workflows I run daily inside Hermes Workspace V2.
Use Case 1 — SEO Research Swarm
I have a profile called "SEO research" with the SEO skills loaded, my keyword tracker file pinned to context, and a memory store with my client domain history.
I chat with the lead agent and it spawns sub-agents for keyword research, SERP analysis, and competitor scraping.
The Kanban board tracks every task and the inspector shows me which agent is doing what.
I cover the broader SEO workflow in Hermes SEO.
Use Case 2 — Content Production
Separate profile called "content writing" with the writing skills loaded.
Different memory store with my voice/style guide, banned phrases, and client tone references.
Agent drafts, I review in the chat panel, the inspector shows me which sources it used.
Use Case 3 — Code Review
"Code review" profile with the developer skills loaded, file panel pointed at the project directory.
Agent reviews PRs, flags issues, suggests refactors.
The terminal panel is still there for running tests.
Use Case 4 — Multi-Agent Orchestration
This is where V2 really shines.
I have a "swarm" profile that runs 5 agents simultaneously — researcher, writer, fact-checker, editor, and publisher.
The virtual office view shows them all working together.
I just feed the swarm a topic and it produces a finished piece.
If you want to see how the swarm works in depth, Hermes Agent Swarm covers it.
Ready to build your own agent swarm? The AI Profit Boardroom has my full multi-agent swarm setup inside the Hermes section, plus the 2-hour Hermes course. Twin guarantee — if I don't double your output in 90 days, full refund. Join 3,000+ members here
V2 Setup Tips Most People Miss
A few things I figured out the hard way.
First, set up your profiles BEFORE you import your old V1 memories.
If you dump everything into one profile you've recreated the V1 mess inside V2.
Second, use the "context budget" setting per profile.
V2 lets you set max tokens per profile, which stops a runaway conversation from blowing your model bill.
Third, pin the inspector panel during development.
You'll catch bugs and weird agent behaviour way faster when you can see reasoning live.
Fourth, the mobile PWA is genuinely useful — install it on your phone and you can chat with your agent from anywhere.
I use it on the train and during walks for quick research questions.
Should You Upgrade To V2 Right Now?
Honest answer based on what I've seen.
If you're on V1 and using it daily — yes, upgrade this week.
The chat experience alone is worth the migration time.
If you're on V0.9 (the old Hermes dashboard) — yes, no question, V2 is a different league.
If you've never used the workspace at all and are running Hermes from the terminal — go straight to V2.
Skip V1 entirely, it's a worse experience and you'd just have to migrate later.
If you've built deep custom integrations with V1 — plan a migration weekend, test in parallel, then cutover.
The migration isn't hard, just don't do it on a deadline day.
V2 Roadmap From What I've Heard
The team has signalled a few things coming next.
Better team/multi-user mode for agencies wanting to share workspaces.
Native integrations with more model providers including local models via Ollama Hermes.
Improved virtual office view with custom agent avatars.
A marketplace for community-contributed skills.
If any of those land, this post gets updated.
V2 vs Other Agent Workspaces
V2 is the strongest open-source agent workspace I've tested.
The closest competitors are OpenClaw's mission control and Claude Desktop, both of which are good but not as flexible.
I've done a full breakdown in Hermes Vs OpenClaw.
The short version: if you want open-source, full agent control, and a proper UI, V2 leads.
If you want polished commercial UX with a smaller feature set, OpenClaw is fine.
If you only need a chat interface, Claude Desktop is enough.
FAQ — Hermes Workspace V2
Is Hermes Workspace V2 free?
Yes, fully free and open-source. You only pay for your AI model usage.
Do I need to uninstall V1 before installing V2?
No, they can run side by side. Test V2 first, then retire V1 when you're confident.
Will my V1 memories transfer to V2?
Yes, V2 reads V1 memory stores automatically. Custom plugins need porting.
Can I run Hermes Workspace V2 on my phone?
Yes, it's a progressive web app — install from the URL on iOS or Android.
Does Hermes Workspace V2 work with local models?
Yes, via Ollama or Atomic Chat — see Hermes Ollama setup.
What's the difference between V2 and Hermes Agent Workspace?
They're the same product — V2 is the current release of Hermes Agent Workspace.
How long does the V2 install take?
5-10 minutes with the Hermes-assisted install. 15-20 if you do it manually.
About Julian
I'm Julian Goldie — AI entrepreneur, SEO expert, and founder of the AI Profit Boardroom (3,000+ members). I help business owners scale with AI agents, automation, and SEO.
- 282K+ YouTube subscribers
- 7-figure AI agency (Goldie Agency)
- Daily training inside the Boardroom
- Author of multiple AI automation playbooks
Get my best AI training inside the AI Profit Boardroom
Also On Our Network
- Read on aiprofitboardroom.com
- Read on juliangoldieaiautomation.com
- Read on aisuccesslabjuliangoldie.com
- Read on aimoneylabjuliangoldie.com
Related Reading
- Hermes Workspace — the original guide to what Hermes Workspace actually is.
- Hermes Agent Workspace — the broader agent workspace overview.
- Hermes Agent Swarm — running multi-agent setups inside V2.
- Hermes Kanban — using the new task board panel.
- Hermes SEO — applying V2 to real SEO workflows.
📺 Video notes + links to the tools 👉
🎥 Learn how I make these videos 👉
🆓 Get a FREE AI Course + Community + 1,000 AI Agents 👉
If you're running Hermes daily, Hermes Workspace V2 is the upgrade that turns your scattered terminal workflow into a proper command centre worth opening every morning.
