How to Deploy and Configure Hermes Agent on QNAP NAS with Container Station

Container Station will automatically create the hermes-data folder on first run when the volume path is declared in the Compose file. No prior folder creation or SSH access is required. Alternatively, the folder can be created in advance via File Station.

1. Open Container Station in the QNAP web interface.
2. Navigate to Applications and click Create.
3. Select Create Application and paste the following Docker Compose YAML:

4. version: "3.8"
   
   services:
     hermes:
       image: nousresearch/hermes-agent:latest
       container_name: hermes
       restart: unless-stopped
       command: gateway run
       volumes:
         - /share/Container/hermes-data:/opt/data
   
   

5. Click Create. Container Station will pull the image and start the container. The hermes-data folder will be created automatically at /share/Container/hermes-data.
6. Verify the container is running by checking the Containers list. The status should show Running.

Note: All credentials are managed by the setup wizard in Step 2 and stored in the data volume. 

After the container is running, open an interactive shell session to complete first-time configuration. The setup wizard configures your LLM provider, API keys, and messaging platform tokens, writing all settings to the persistent data volume.

1. In Container Station, select the hermes container.
2. Click Execute and select /bin/bash (or /bin/sh if bash is unavailable) to open an interactive shell.
3. In the shell, run:
   hermes setup

4. 

   

5. Follow the on-screen prompts to configure:
   • LLM provider and model (e.g., OpenAI, Anthropic, OpenAI Codex)
   • API keys
   • Messaging platform (e.g., Telegram bot token) — recommended at this stage
6. The wizard writes your settings to /opt/data/config.yaml and /opt/data/.env. This step only needs to be performed once.
7. After the wizard completes, type exit to close the shell, then restart the container via Container Station to apply the configuration.

1. In Container Station, select the hermes container and click Logs.
2. Confirm the following lines appear in the output:
   Hermes Gateway Starting...
3. Click Execute and select /bin/bash (or /bin/sh if bash is unavailable) to open an interactive shell.
4. In the shell, run:  hermes 
5. Then you can talk to it.
6. If using Telegram, send a message to your configured bot to verify the connection.

Use the hermes model command to interactively switch providers or models at any time.

1. In Container Station, select the hermes container.
2. Click Execute and select /bin/bash (or /bin/sh ).
3. In the terminal session, run:
   hermes model
4. Follow the interactive prompts to select a new provider and model. The selection is saved to config.yaml automatically.

Supported providers include: anthropic, openai-codex, openai, openrouter, nous, deepseek, minimax, and others. For a full list, refer to the official CLI reference.

Use hermes setup gateway to add or update messaging integrations (Telegram, Discord, Slack, WhatsApp, Signal, etc.) without re-running the full setup wizard.

1. In Container Station, select the hermes container.
2. Click Execute and select /bin/bash.
3. Run:
   hermes setup gateway
4. Follow the prompts to add or update your messaging platform credentials.
5. Restart the container after completing the wizard:
   exit
   Then restart via Container Station — select the container and click Restart.

To reconfigure any part of the setup (model, terminal backend, tools, gateway, agent personality), run the full wizard or a specific section:

1. Open a shell in the container via Container Station Execute.
2. Run the full wizard:
   hermes setup
3. Or run a specific section only:
   • hermes setup model — LLM provider and model
   • hermes setup gateway — Messaging platform configuration
   • hermes setup tools — Enable or disable tools
   • hermes setup terminal — Terminal backend (local, Docker sandbox, SSH)
   • hermes setup agent — Agent identity and personality (SOUL.md)
4. Restart the container after making changes.