Hermes Studio

What Hermes Studio does

Hermes Studio connects directly to the Hermes Agent gateway with live SSE streaming and renders tool calls in the chat. Its multi-agent crews let you create named groups of agents, dispatch tasks to all or specific members and watch live activity feeds, with per-profile workspace roots giving each member an isolated file view. A visual workflow builder runs DAG-structured task pipelines for crews.

A cron job manager lets you create, edit, pause, trigger and monitor scheduled jobs, with manual runs streaming tool events into the job card. You can approve, deny or always-allow shell commands, browse and edit agent memory as a force-directed knowledge graph of wiki links, and install, uninstall or toggle skills from the browser. Settings cover approvals, command allowlists, toolsets and a security scanner. The UI has eight themes and is a mobile-friendly PWA reachable through Tailscale, and it works with Ollama, OpenAI, Anthropic and other OpenAI-compatible backends.

Key features

• Multi-agent crews with task dispatch, live activity feeds and profile-scoped workspaces
• Cron job manager to create, pause, trigger and monitor scheduled Hermes jobs
• Execution approvals for agent shell commands, with always-allow option
• Memory browser with an interactive knowledge graph, plus skill install and toggle
• Visual DAG workflow builder for crew pipelines
• PWA with eight themes, Redis-backed session persistence and Docker and Tailscale guides

When to use it

• Scheduling and monitoring recurring Hermes tasks without touching a terminal
• Approving or denying agent shell commands from a browser or phone
• Running a small crew of Hermes agents and watching what each member does

Who it is for: Hermes Agent users who prefer a browser workspace for running agents locally.

How it fits with Hermes Agent

It is a dedicated web UI for Hermes Agent that talks to the Hermes gateway, and the README compares itself with the hermes-workspace UI.

FAQ