Hermes-Agent

Hermes Agent

It’s an open-source AI agent created by Nous Research, designed to be a persistent, self-improving personal assistant. It has 8.7k on GitHub. Here are the key points:

It’s the only agent with a built-in learning loop: it creates skills from experience, improves them through use, remembers past conversations and builds a deep model of who you are across sessions.

Main features

Persistent memory — Remembers who you are between conversations. It doesn’t start from scratch every time.

Skills — After completing complex tasks, the agent automatically creates reusable “skills” and keeps improving them over time.

Multi-platform — You can talk to it from Telegram, Discord, Slack, WhatsApp, Signal and the CLI, all from a single gateway process.

Runs anywhere — A $5 VPS, a GPU cluster, or serverless infrastructure that costs almost nothing when idle. It isn’t tied to your laptop.

Any AI model — You can use OpenRouter (200+ models), OpenAI, your own endpoint, etc. You switch with hermes model without touching code.

Scheduled automations — Built-in cron: daily reports, nightly backups, all in natural language.

Parallel subagents — It can delegate tasks to subagents that work in parallel.

In short, it’s like a personal AI assistant that lives on your server, that you can use from your phone via Telegram while it works in the cloud, and that becomes more useful for you specifically over time. It’s aimed at both technical users and AI researchers.