sharbelxyz/hermes-agent-mission-control

What it solves

Hermy HQ provides a centralized, self-hostable dashboard for managing a local Hermes AI agent. It eliminates the need to interact with the agent via a command line, allowing users to dispatch tasks, monitor activity, and manage the agent's memory through a visual interface.

How it works

The project uses a "message bus" architecture where a Next.js website and a local bridge process communicate via a shared PostgreSQL database. The website sends requests to the database; a small bridge process running on the same machine as the Hermes agent polls this database, executes the commands via the hermes CLI, and writes the results back to the database for the website to display.

Who it’s for

Users who have a self-hosted Hermes agent and want a professional "cockpit" to manage their AI agent's workflows, memory, and task approvals without exposing their local machine to the internet.

Highlights

• Approval Inbox: Prevents the agent from performing unexpected actions by requiring manual approval for any request with side effects.
• Memory Wiki: A visual interface to browse and edit the facts, notes, and links stored in the agent's memory.
• Command Palette: A ⌘K shortcut for quick navigation and action triggering.
• Content OS: Built-in pipelines for managing X/Twitter content, YouTube scripts, and longform writing.
• Secure Architecture: The bridge only requires outbound access to the database, meaning no inbound ports need to be opened on the local machine.