Hermes Agent: From Install to Advanced Automation

Understand the difference between an autonomous agent, a chatbot, and a coding copilot, and discover what makes Hermes uniquely self-improving through its skills, memory, and provider-agnostic architecture.

Explore the real-world applications of Hermes — from software development and research to system administration and multi-step automation — so you can identify where it fits into your own work.

Confirm your platform is supported (macOS, Linux, or Windows/WSL), review the LLM providers Hermes works with, and set the stage for a smooth installation

Run `hermes setup` to complete the initial configuration wizard and take the fastest path to a working agent using `hermes setup --portal` with Nous Portal OAuth.

Link Hermes to your preferred AI provider — Nous Portal, Anthropic, OpenAI, DeepSeek, OpenRouter, or a local model — by placing API keys in `~/.hermes/.env` and selecting a model interactively.

Run `hermes doctor` for a full health check, use `hermes status` to confirm each component is live, and send your first query with `hermes chat -q "Hello, Hermes!"`.

Learn where Hermes stores its config (`~/.hermes/config.yaml`), secrets (`.env`), session database (`state.db`), and skill library so you can confidently navigate and manage your installation.

Understand how Hermes reasons through tasks, selects and calls tools, and constructs its responses — giving you the mental model you need to predict and guide its behaviour

Learn how toolsets extend Hermes's capabilities — covering file access, web browsing, code execution, and more — and how to control which ones are active for any given session

Understand the skills system as Hermes's procedural memory, see how it powers the self-improvement loop, and explore agentskills.io as a community hub for sharing and discovering skills.

Launch Hermes in interactive or one-shot mode, and learn how to pick up where you left off with the `--continue` and `--resume SESSION` flags.

Master the in-chat commands that let you control sessions (`/new`, `/retry`, `/undo`), switch models (`/model`), view history, save outputs, and adjust agent behaviour on the fly.

Use the `file`, `terminal`, and `code_execution` toolsets to read, write, search, and patch files, run shell commands safely, and let Hermes operate directly in your codebase

Put Hermes to work on the internet with web search, page extraction, and browser automation, plus vision analysis and image generation for multimedia tasks.

Learn when and how to compress context manually with `/compress`, understand auto-compression thresholds, and use `/rollback` to restore filesystem checkpoints when things go wrong.

View your current configuration with `hermes config`, open it in your editor with `hermes config edit`, and set individual values programmatically with `hermes config set KEY VALU

Walk through every major config section — model, agent, terminal, display, compression, memory, and security — and understand precisely what each setting controls and when to change it.

Use the `hermes tools` curses UI or simple CLI subcommands to enable and disable toolsets, and understand why a `/reset` is sometimes needed for changes to take effect.

Store credentials securely with `hermes auth`, set up credential pools for automatic API key rotation, and manage multiple provider keys across environments

Understand skills as SKILL.md procedural memory files that persist expert workflows across sessions, and see how they form the foundation of Hermes's self-improvement loop.

Explore the skill catalog on agentskills.io, search and list your installed skills, and load a skill for the current session using `/skill name` or the `-s` flag.

Install skills from the hub or a direct URL, preview them with `hermes skills inspect` before committing, and keep them current with `hermes skills update`.

Write a SKILL.md file with the correct frontmatter and markdown structure, learn when to capture a skill after complex iterative tasks, and publish it to agentskills.io.

Let Hermes automatically manage the skill lifecycle — pinning frequently used skills, archiving stale ones, and rolling back bad versions — all accessible via the `/curator` slash command.

Understand the difference between user profile memory and agent notes, what Hermes saves automatically versus what requires explicit action, and how FTS5 full-text search enables cross-session recall.

Compare the built-in memory system with Honcho and Mem0 integrations, and configure your preferred provider with `hermes memory setup` and `hermes memory status`.

Find old conversations and extract information from them using `hermes sessions browse` and the `session_search` tool available from inside a live session.

Learn what to save versus what to leave ephemeral, and how to keep your memory store compact, accurate, and immediately actionable across long-running projects.

Discover how the Gateway lets you interact with your full-featured Hermes agent — including all toolsets — from over 20 messaging platforms without losing any capability.

Use the `hermes gateway setup` interactive wizard to configure platforms, run the gateway in the foreground for testing, and install it as a persistent background service.

Step-by-step setup for the most popular platforms, including bot tokens, Message Content intents for Discord, event subscriptions for Slack, and email/SMS/WhatsApp configuration.

Control your gateway sessions from any connected platform using commands like `/approve`, `/deny`, `/restart`, and `/sethome` to manage behaviour without touching the terminal.

Diagnose gateway failures by reading `~/.hermes/logs/gateway.log`, fix lingering process and WSL2 systemd issues, and handle the most common platform-specific gotchas

Create, edit, pause, and run scheduled Hermes jobs using human-friendly intervals or cron expressions, with per-job model overrides, skill loading, workdir settings, and job chaining via `context_from`.

Subscribe to named webhook events, trigger full Hermes agent runs from external services (GitHub, CI/CD, monitoring tools), and validate your setup with `hermes webhook test`.