Hermes AgentYour agent, your hardware

A personal handbook for running Hermes Agent fully offline with open-source models on your laptop or local server. Memory, skills, tools, and real workflows that compound over time.

Start here

Offline setup

Use cases

🔒

Fully offline

Run open-weight models via Ollama on your laptop. No API keys, no data leaving your machine.

Set up offline

🧠

Self-improving agent

Hermes builds skills from experience, persists memory across sessions, and learns your preferences over time.

The learning loop

🔧

60+ built-in tools

File editing, shell commands, web tools, delegation, and MCP. An agent that acts, not just chats.

Explore tools

📱

Talk from anywhere

CLI, Telegram, Discord, Slack, and 20+ platforms. Your local agent, reachable from your phone.

Messaging setup

⚡

Real use cases

Practical offline workflows: file assistant, private research, code helper, daily briefings.

Browse use cases

⏰

Scheduled automations

Built-in cron to run tasks on a schedule and deliver results to any connected platform.

Set up cron

Who this is for

This is my personal learning path for Hermes Agent, the self-improving agent built by Nous Research. I run it in complete offline mode on my local machine with open-source models via Ollama.

By the end you'll know how to:

Install Hermes and connect it to a local model server.
Configure context length, tool calling, and performance for agentic work.
Use memory and skills so the agent gets better the longer it runs.
Apply real offline workflows to everyday tasks.
Optionally extend with messaging gateways and scheduled jobs.

How to read it

If you want to...	Go to
Understand what Hermes is	Meet Hermes Agent
Run everything offline	Complete offline setup
Pick the right local model	Choosing a model
Solve a specific task	Real use cases
Keep the agent safe	Security model
Look up a command	Command reference

A note on freshness

Hermes Agent evolves quickly. This handbook reflects my setup as of mid-2026. When in doubt, check the official docs.