Hermes Agent

Best for: people who want a persistent AI assistant that accumulates context across sessions instead of starting cold every time.

Hermes Agent, built by Nous Research, is a self-improving agent with a built-in learning loop. It creates skills from experience and refines them during use, prompts itself to persist knowledge, searches its own past conversations, and builds a model of the person it works with over time. It is model-agnostic — Nous Portal, OpenRouter, OpenAI or your own endpoint all work.

What makes it useful: it is not tied to your laptop. Hermes runs on a $5 VPS, a GPU cluster, or serverless infrastructure that costs almost nothing when idle, and you can talk to it from Telegram while it works on a cloud VM. The memory across sessions is the real differentiator.

Where it falls short: self-hosting means you own deployment, model costs and security. An agent that remembers everything is also an agent accumulating a detailed record of your work, so think about where that data lives before you commit. Setup expects comfort with servers and API keys.

Details: Python, MIT licensed. Bring your own model.