Video: "Hermes Workspace: New Mission Control is INSANE!" by Julian Goldie on YouTube.
What the old setup actually felt like
Hermes Agent V1 worked well if you were comfortable in a terminal. You ran commands, got output, managed memory through config files, and handled scheduling via cron. For developers that was fine. For anyone else — a business owner, a content manager, a small team without a dedicated engineer — it was a meaningful barrier. You also had to drop out to Telegram or a terminal window just to have a conversation with your agent.
The result was that Hermes stayed squarely in the hands of technical users, even though the underlying capability was genuinely broad. V2 is the attempt to fix that.
What V2 actually adds
The centrepiece is browser chat. You open Hermes Workspace in your browser and talk to your agent directly, without touching a command line. Sessions are logged, history is searchable, and you can switch between conversations while keeping each one in context. That sounds basic but it removes the single most common reason people abandoned Hermes after installation.
Alongside that: a memory browser that lets you read and edit what your agent knows, a skills manager for adding or disabling capabilities without editing config files, and an inspector that shows you tool calls and model switches as they happen. Worth noting — the inspector is the closest thing Hermes has to a built-in equivalent of Labyrinth. It is not as detailed, but it covers the basics without a separate plugin.
Multi-profile management — why that matters
Each profile in V2 gets its own system prompt, its own model selection, its own skill set, and its own memory layer. You flip between them with one click. In practice this means you can have one profile set up for SEO research, a second for coding tasks, a third for client communication, and they do not bleed into each other.
Previously you needed separate Hermes installations or a lot of manual config switching to achieve the same thing. For anyone running Hermes for more than one type of work — which is most businesses — this is probably the most useful change in V2.
The Kanban board and mobile angle
V2 adds a Kanban task board for managing agent work in progress, which pairs with the existing swarm mode. If you are running specialist agents in parallel — a researcher, a writer, a reviewer — you can now see where each one is and what it is blocking on. It is a visual layer on top of what was previously invisible background work.
The PWA support means V2 installs as an app on iOS and Android. Not the same as a proper native app, but functional for checking on an agent run or giving it a new brief while you are away from a laptop.
The honest caveats
V2 is a structural rebuild rather than a polished product. Early reports suggest some rough edges around session persistence and skill loading order. The terminal is still there for anything the GUI does not expose — which at launch is quite a lot of the advanced configuration. If you are running a production Hermes setup for a client, wait for a point release before migrating. For personal use or evaluation, it is worth setting up now.
Where this connects to NordSys
We install and configure Hermes Agent for clients — which increasingly means setting up Workspace V2 as the daily interface, not just the raw agent. If you want Hermes running properly with the right profiles, the right models, and the right skills for your specific workflow, that is what we do. We also keep it updated as the project moves quickly.
See our AI Agents service →