How we built the Hermes OS dashboard — and open-sourced the whole thing

When you're running 12+ cron jobs across multiple platforms, you need a way to see what's actually happening.

Not logs in a terminal. Not Discord notifications buried in a channel. A proper dashboard — something that shows you the state of every job, every activity, every task, in one place.

That's why we built Hermes OS.

And instead of keeping it to ourselves, we put the entire thing on GitHub for anyone to use, fork, or improve.

---

What is Hermes OS?

Hermes OS is a mission control dashboard for Hermes AI agent. Think of it as the control room for your agent operations — every cron job, every activity log, every task, all in one interface.

Here's what it gives you:

Overview — a real-time snapshot of all active jobs, system health, and recent activity. At a glance, you know what's running, what's queued, and what needs attention.

Chat — talk directly to your Hermes agent through the dashboard. Same agent that's running your automations, just through a clean web interface instead of Discord or Telegram.

Activity — a chronological feed of everything your agent has done. Lead captures, uptime checks, idea generation sessions, content drafts — all logged and searchable.

Tasks — create and manage tasks for your agent to execute. Assign priorities, track completion, and keep work organized.

Models — see which AI models are being used, monitor usage, and switch between providers when needed.

Config — manage your agent's configuration from the dashboard. No more editing config files in a terminal.

Logs — full system logs with filtering and search. When something breaks, you can find out why fast.

It's the same way we monitor our own Hermes setup — the one running Comfort Shooting's operations, generating ideas at 3 AM, monitoring websites, and handling lead capture in real-time.

---

Why we built it

We started with cron logs in Discord. That worked for a while. But as the number of jobs grew — real-time lead monitoring, uptime checks, trading briefings, nightly idea generation, content pipeline, more content ideas being generated — Discord became noise. Important signals got buried in the chatter.

We needed:

1. A single source of truth — one place to see everything, not scattered across Discord channels and terminal logs
2. Real-time status — know immediately if a job failed, not 3 hours later when you happen to check Discord
3. Config management — adjust settings without SSH-ing into a server
4. Chat interface — talk to the agent without context-switching to another platform

So we built it.

---

The stack

Hermes OS is built with:

• Next.js 15 — App Router, React 19, static + dynamic rendering where it makes sense
• Supabase — PostgreSQL database for real-time activity logs, job state, and config
• Tailwind CSS v4 — utility-first styling with a custom design system (light/dark mode, glassmorphism, animated backgrounds)
• Vercel — deployed and auto-deploying on every push to main
• Hermes API — a custom API server running on port 8642, tunneled via Cloudflare, that the dashboard talks to for chat, cron control, and live data

The whole architecture is deliberately simple:

Hermes Agent (port 8642)
        ↕
Hermes OS Dashboard (Next.js 15 + Supabase)
        ↕
Vercel (auto-deploy)

The dashboard doesn't replace the agent — it wraps it. Your agent keeps doing its thing. The dashboard just gives you visibility and control.

---

Why it's open-source

We didn't build Hermes OS just for us. We built it because if you're running Hermes at any serious level, you need something like this.

The repo is at github.com/Mr-Akhil12/hermes-mission-control. Public. MIT-licensed. Fork it, deploy your own, add features, customize it for your setup.

Here's why we open-source everything we build:

Because gatekeeping doesn't scale. If someone in Nairobi or Lagos or London can take our dashboard, modify it for their setup, and run their own agent operations more effectively — that's a win. We're not in competition with people using our tools. We're in the business of making AI agents accessible.

Because transparency builds trust. When you can see exactly how something works — the code, the database schema, the deployment config — you can verify it does what it claims. No black boxes. No "trust us." Here's the code. Read it.

Because the best features come from other people. Someone will look at Hermes OS and think "I wish it did X." And then they'll build X and open a pull request. That's how open-source works. That's how we all get better.

---

What we're working on next

Hermes OS is functional but it's not finished. The roadmap includes:

• Notifications — push notifications for job failures, not just in-dashboard alerts
• Mobile app — a proper mobile interface, not just a responsive website
• Multi-agent support — monitor multiple Hermes instances from one dashboard
• Analytics — job success rates, execution times, resource usage over time
• Plugin system — let other developers build custom dashboard widgets

If you want to contribute, the repo is open. If you want to use it as-is, clone it, configure your Supabase, point it at your Hermes API, and deploy.

---

The point

We build tools. We use them. We open-source them.

That's the whole model. We're not sitting on a proprietary dashboard charging $99/month. We built something useful, we use it every day to run our own agent operations, and we're giving it away.

Because the real value isn't in the dashboard. It's in what you do with it.

---

Hermes OS is live at jarvis-x-akhil.vercel.app. The source code is at github.com/Mr-Akhil12/hermes-mission-control. Fork it, use it, improve it.

If you want help setting up your own Hermes agent or deploying the dashboard for your team — book a free consultation. We don't gatekeep, but we will help you build.