Hermes Agent

Autonomous AI agent from Nous Research, with chat in the browser and messaging channels you connect when you want them.

Overview

This template runs Hermes Agent by Nous Research. Hermes is an autonomous agent: you give it a task and it works on it with its own tools rather than answering a single prompt. The template wraps the upstream image with a start command (upstream's default process is an interactive terminal UI that exits without a TTY) and runs the web dashboard as the main process with the messaging gateway supervised beside it, state on a persistent volume.

The dashboard is what the service URL serves. Chat lives there too, so a fresh deploy is usable with no messaging platform configured at all. When you want the agent on Telegram, Slack, Discord, WhatsApp, or any other platform the gateway supports, the dashboard's Channels page connects it after deploy: a per-platform form, an enable toggle, and a restart button that applies the change.

What you get by hosting it

• An HTTPS URL for the Hermes dashboard, behind a username and password you choose, with no port forwarding or tunnel to manage. Upstream refuses to expose the dashboard without one.
• Chat with the agent directly on the dashboard, no bot token required.
• The Channels page: connect any supported messaging platform (Telegram, Slack, Discord, WhatsApp, Signal, Email, and more) from the browser, with connection status and a gateway restart button, so channel setup never needs a redeploy.
• A persistent volume mounted at /data, with HERMES_HOME pointed at /data/.hermes so the agent's state, and every channel you configure, survives restarts and redeploys.
• The gateway token generated for you (64 chars) and stored as a managed secret, so it is never committed anywhere or shown in the manifest.
• Your API keys and any bot tokens held as managed secrets rather than baked into an image.
• Deploys are health-gated against /api/status, so a container that never becomes ready is rolled back rather than left serving errors.

What you need before deploying

• A username and a password of your choosing for the dashboard sign-in. There is no default and nothing is generated: the deploy form starts with both fields empty and will not submit until you fill them. A password the platform minted would be one it could never show you again.

That is all. Everything else is optional and configurable from the dashboard after deploy: an OpenRouter API key on the API Keys page (the agent calls models through OpenRouter, so chat starts answering once one is set; fill the variable at deploy time to skip that step), and messaging platforms on the Channels page.

Configuration

Set by the template, not by you: HERMES_HOME=/data/.hermes (state on the volume) and HERMES_DASHBOARD_PORT=8080. The entrypoint binds the dashboard and starts the gateway as a managed daemon the dashboard's System and Channels pages control.

After deploy

1. Wait for the health check on /api/status to pass, then open the service URL and sign in with the ADMIN_USERNAME and ADMIN_PASSWORD you deployed with.
2. If you left OPENROUTER_API_KEY blank, open the API Keys page and set it there first.
3. Give the agent a task in the dashboard's Chat tab. This works immediately, with no messaging platform configured.
4. To reach the agent from a messaging app, open the Channels page: pick the platform, fill its form (each shows exactly the fields that platform needs and links its credential docs), enable it, and restart the gateway from the same page. Tokens land on the volume, so they survive restarts and redeploys.
5. Platforms deny unknown senders by default. Add your user ID to the platform's allowlist, or approve yourself via the pairing flow when the bot replies with a pairing code.
6. State lives under /data/.hermes, so restarts and redeploys keep the agent's memory. Deleting the volume resets it.

Links

• Architectures: linux/amd64 and linux/arm64. Upstream publishes both and this image only adds an entrypoint.
• Documentation: https://hermes-agent.nousresearch.com/docs/
• Upstream: https://github.com/NousResearch/hermes-agent
• Image: docker.io/nousresearch/hermes-agent, pinned to v2026.8.27
• License: MIT (upstream NousResearch/hermes-agent).