The hermes live dock is the new CLI and TUI panel in Hermes Agent v0.21.5 that keeps your standing /goal and queued prompts on screen while the agent works — released on 24 September 2026, it means you no longer have to interrupt a running session to remember what your agent is chasing or what is stacked up next. For anyone running Hermes as an automation workhorse rather than a chat toy, that is a genuinely useful change: long-running goal loops finally come with a permanent, glanceable status line. This guide covers what the live dock shows, how it fits the goals system it sits on top of, and how to fold it into a working automation routine.
📺 Watch: Hermes AI Agent Now Has Its Own Computer
🔥 Get the Agent OS as a free bonus: AI Profit Boardroom members get the full Agent OS zip, prompt libraries, daily tutorials and weekly live coaching calls. → Get inside · Want AI SEO help 1-on-1? Book a free SEO strategy session →
The sourcing first: the live dock ships in Hermes Agent v0.21.5, published by Nous Research on GitHub on 24 September 2026 under the tag v2026.9.24. The release notes describe it as a live dock displaying standing /goal and queued prompts in the CLI and TUI — one line in a patch that rolls up roughly 460 pull requests, but the line that matters most if your Hermes setup runs unattended work.
What the Hermes Live Dock Shows You
Two things, per the v0.21.5 release notes: your standing /goal, and your queued prompts. If you have not used goals yet, /goal is the Hermes command that hands the agent a persistent objective and lets a native loop drive it turn after turn until a judge model decides the goal is complete — the full mechanics are in the Hermes agent goals guide. Queued prompts are exactly what they sound like: instructions stacked up behind whatever the agent is currently doing.
Before the live dock, both of those things were invisible unless you went looking. A goal loop might be twenty turns into a fifty-turn budget with three prompts queued behind it, and the terminal showed you only the current turn's output scrolling past. The dock pins the missing context to the interface: what the agent is working towards, and what is waiting. It is the difference between watching a worker's hands and seeing their task list on the wall behind them.
If you want long-running agent workflows that earn while you sleep — goals, queues and the full automation stack — the AI Profit Boardroom has the templates and weekly live coaching calls to build it with you → Set up your goal-driven agents inside. Want 1-on-1 AI SEO help instead? Book a free SEO strategy session.
Why the Live Dock Matters for Automation Workflows
The live dock is a visibility feature, and visibility is what makes autonomous loops trustworthy enough to actually use. The goals system already handled the hard parts — the accept-execute-judge loop, budgets via goal_max_turns, completion detection via a configured judge model, pause and resume across sessions. What it lacked in the terminal was a persistent answer to the operator's two standing questions: what is this agent doing, and what happens next?
That has practical consequences for how you run automation:
- You can queue work with confidence. Stacking prompts behind a running goal only makes sense if you can see the stack. The dock shows the queue, so batching your day's tasks into one session stops feeling like throwing notes over a wall.
- You catch drift early. A standing goal pinned on screen is a constant check that the loop is still pointed where you aimed it, without waiting for the final judge verdict.
- Long sessions become reviewable at a glance. Coming back to a terminal after an hour, the dock answers "where are we?" before you scroll a single line of output.
If you monitor several bots rather than one session, the dock complements rather than replaces the bigger surfaces: the Hermes agent mission control view for fleet-level oversight, and the Hermes kanban board for task-level tracking. The dock is the in-terminal layer of the same idea — status where you are already looking.
Using the Hermes Live Dock Day to Day
A sensible working pattern with the live dock looks like this. Update Hermes to v0.21.5 first — the dock ships with the release, so hermes update is step zero. Then run your normal goal workflow: set a /goal with a completion condition you would trust a judge model to verify, keep goal_max_turns proportionate to the job (the goals guide suggests 20 to 50 depending on complexity), and queue follow-up prompts behind it rather than waiting for the loop to finish.
With the dock on screen, the queue becomes a planning surface. Front-load the session: goal first, then the prompts you know you will want afterwards — the summary, the follow-up checks, the handoff notes. The dock keeps the whole plan visible in the CLI and TUI, so mid-session you are managing a visible pipeline rather than reconstructing intent from scrollback. For scripted pipelines that consume Hermes output programmatically, pair the dock with Hermes agent stream-json output — the dock is for your eyes, the JSONL stream is for your code, and mature automation setups use both.
A concrete example makes the shape obvious. Suppose the standing goal is a content-refresh sweep: a /goal telling the agent to update every post in a folder until a checker pass finds no stale figures. Behind it you queue three prompts — a summary of what changed, a list of pages needing a human eye, and a draft handover note. The live dock keeps all four items in view for the whole run: the goal up top, the queued jobs beneath it. Nothing about that workflow was impossible before v0.21.5; it was simply invisible, and invisible workflows are the ones you quietly stop trusting.
One planning note: the dock shows standing goals and queued prompts, so it rewards setups that actually use goals. If your Hermes usage is still one-shot prompting in the Hermes Claude CLI style, the dock will not have much to show — which is itself a useful signal about how much of the automation surface you are leaving unused.
Hermes Live Dock at a Glance
| Question | Answer |
|---|---|
| What is it? | A live panel showing your standing /goal and queued prompts |
| Where does it appear? | The Hermes CLI and TUI |
| Which version added it? | Hermes Agent v0.21.5 (tag v2026.9.24), released 24 September 2026 |
| What does it depend on? | The existing goals system (/goal, goal_max_turns, judge models) and prompt queueing |
| What is it for? | Keeping long-running, queued agent work visible without interrupting the session |
📺 Watch: Automate Anything using the NEW AgentOS System!
Hermes Live Dock FAQs
What is the live dock in Hermes Agent?
A panel added in v0.21.5 that displays your standing /goal and queued prompts in the CLI and TUI, so long-running agent sessions carry their own always-visible status instead of burying it in scrollback.
Do I need to configure anything to get the live dock?
The release notes list it as a v0.21.5 CLI and TUI feature rather than an opt-in setting, so updating Hermes to v0.21.5 is the requirement. It becomes useful the moment you run a /goal or queue prompts.
Is the live dock the same as the Hermes dashboard?
No. The Hermes dashboard command opens a full visual dashboard with voice and build panels, while the live dock is a lightweight status panel inside the terminal interfaces. They answer different scales of the same question.
Does the live dock work with bot rosters?
The release notes describe it as a CLI and TUI feature, so it is per-session. For roster-wide visibility across multiple bots, mission control remains the right surface.
Verdict: Small Panel, Big Habit Change
The Hermes live dock is the kind of release-note one-liner that quietly changes daily behaviour. Goals made Hermes capable of long, unattended work; the dock makes that work legible while it happens, and legibility is what lets you delegate more without anxiety. Update to v0.21.5, set a real goal with a queue behind it, and let the dock do the remembering. For where this fits in the bigger machine — orchestration, memory and mission control as one system — the Agent OS guide is the map, and the Goldie Bench write-up covers how the models you might run behind these loops compare in hands-on tests.
If you want an agent that runs your workflows on goals and queues while you focus on the work only you can do, check out the AI Profit Boardroom → Automate your first goal loop with the Boardroom. And if you want a personal AI SEO plan for your business, book a free SEO strategy session.
