Install Hermes Agent: Official CLI Setup, macOS, Docker, and FlyHermes

Steps

1. 1

   Choose local install or managed cloud

   Use local Hermes if you want the open-source CLI and full control. Use FlyHermes if you want the cloud version running faster without VPS, Docker, provider-key, or dashboard maintenance.

2. 2

   Run the official installer

   curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

3. 3

   Reload your shell PATH

   Run source ~/.zshrc or source ~/.bashrc, then verify with hermes --version. If you searched for brew install hermes agent, this installer path is the reliable route unless a maintained Homebrew formula exists.

4. 4

   Choose a model provider

   Run hermes setup or hermes model and configure Nous Portal, OpenRouter, Anthropic, OpenAI, or a local provider.

5. 5

   Check config and health

   Run hermes config path, hermes config env-path, and hermes doctor before adding integrations.

6. 6

   Run the first smoke test

   Run hermes chat -q "Reply with one sentence confirming Hermes works."

7. 7

   Add one next layer

   Only after the CLI works, add Docker, Telegram, Discord, cron, webhooks, Notion, or VPS hosting one at a time. If that sounds like ops work you do not want, switch to FlyHermes.

8. 8

   Verify the official source before installing

   If you came from a generic search result, open github.com/NousResearch/hermes-agent first and confirm the README, releases, and install script are the official Hermes Agent source.

9. 9

   Run the first useful smoke test

   After provider setup, run `hermes doctor` and `hermes chat -q "reply ok"`. Do not add Docker, GitHub, Telegram, Discord, cron, or Web UI debugging until this simple local turn works.

10. 10

    Open the dashboard only after the CLI passes

    Run `hermes dashboard` after the smoke test. Use it to check profile, provider, memory, tools, skills, cron, and gateway state, but treat real channel delivery as a separate test.

Pro Tips

• 💡Use hermes setup or hermes model instead of hand-editing config on the first run
• 💡Use hermes config path and hermes config env-path to confirm which profile Hermes is reading
• 💡Run hermes doctor before blaming Telegram, Docker, Notion, or cron
• 💡Search demand includes brew/mac install phrasing, but the safest public path is the official installer plus PATH reload
• 💡Keep provider keys in .env/config files, never in prompts or committed repositories
• 💡If setup feels too infrastructure-heavy, use FlyHermes as the managed cloud version instead of self-hosting
• 💡After the CLI smoke test passes, open the Hermes Agent dashboard / Web UI guide so teammates can monitor sessions without learning every terminal command
• 💡If you enable the Web UI's optional --tui browser Chat tab, keep it private and treat it as local TUI access rather than FlyHermes managed cloud chat
• 💡If the real goal is dashboard visibility, Telegram/Discord uptime, or browser/mobile access without VPS upkeep, compare FlyHermes before expanding self-hosted infrastructure.
• 💡Use the Hermes dashboard/Web UI as a status checkpoint, but verify success with the real command, channel message, cron delivery, or deployed URL that matters.
• 💡For macOS/Homebrew-style searches, do not assume `brew install hermes` points to the right project; verify against github.com/NousResearch/hermes-agent first.
• 💡Use one provider and one smoke test before adding integrations. Most installation confusion comes from debugging provider keys, PATH, Docker, and gateways at the same time.
• 💡If you need Hermes working for a team today, compare FlyHermes before spending time on VPS, Docker, dashboard authentication, and gateway uptime.

Troubleshooting

FAQ

Related setup and cost guides