summree
Hermes Agent + Ollama = 100% Private OS
Ollama
Jack Roberts

Hermes Agent + Ollama = 100% Private OS

⏱ 19 min video · 3 min read5 Jun 2026Worth watching
TL;DR
Jack Roberts walks through setting up Hermes Agent to run 100% locally and privately using Ollama, so you can use an AI operating system on your own machine at zero cost with no internet required. The video covers model selection, installation steps, context window requirements, and when to use local vs. cloud AI.
Key points
1
Ollama is used to download and run open-source models (Qwen, Deepseek, Gemma, Mistral) locally — free forever, no data leaves your machine
2
Hermes Agent requires a local model with at least 64,000 tokens of context window; Qwen 3 Coder 30B (64K) is the recommended model for Hermes compatibility
3
The best local models today are roughly one year behind frontier models — equivalent in quality to approximately Claude Sonnet 4 as of mid-2025
4
A new Hermes desktop app has been released, providing a less intimidating GUI alternative to the terminal for managing sessions and models
5
The video introduces a 'vault mode vs. connected mode' framework: use local/private for sensitive data, use cloud models when raw performance matters most
Actionable insights
Go to ollama.com, download Ollama, then run 'ollama pull qwen3-coder:30b-a3b-q8_0' (or equivalent 64K model) in Terminal to get a Hermes-compatible local model
Check your MacBook specs via Apple menu > About This Mac, screenshot it, and ask Hermes which Ollama model best fits your hardware before downloading
Use local/vault mode for client data, health notes, proprietary code, and offline work; switch to cloud models when you need the fastest, highest-quality answers
Notable quotes

We had this move, everyone going to the cloud. Now the cloud is old. Now we are going local. Local is the future.

The best local model today is about one year behind wherever we are currently at — so for example, the best local model today is as good as the best model that existed in around mid-2025. That would be Claude Sonnet 4.

I was flying from Dubai to LA and the internet at one point was not working and I was just using it on my computer. It felt really freaking cool.

Worth watching?
Worth watching the full video?
Watch if you want the step-by-step terminal commands and model recommendations — the key facts are all here, but the live screen walkthrough may help if you are new to Ollama or the Hermes desktop app.
Topics
AI & TechOllama

Explore more summaries on these topics →

Saved you some time? The creator still deserves a like.

Watch on YouTube →
More like this

Want this for your own channels?

Add the channels you follow. Every new video summarised and in your inbox the moment it drops. From £4/month.

Try it free