What it does
Hermes Agent is an open-source AI agent built by Nous Research. Its README says it creates skills from experience after complex tasks, improves them as it is used, keeps agent-curated memory, and searches past conversations with summaries to recall earlier work. It is not tied to a particular model: you pick a provider such as Nous Portal, OpenRouter, OpenAI or your own endpoint and switch with a command. You can talk to it in a full terminal interface or through a single gateway process that connects Telegram, Discord, Slack, WhatsApp and Signal. It has a built-in scheduler for unattended jobs delivered to those platforms, can spawn subagents for parallel work, and can connect MCP servers. Seven terminal backends are listed: local, Docker, SSH, Singularity, Modal, Daytona and Vercel Sandbox. Install is a one-line shell command on Linux, macOS and WSL2, a PowerShell command on native Windows, or an APT repository for Android with Termux. The README also describes batch trajectory generation for researchers and an import path for people coming from OpenClaw. A separate Nous Portal subscription is offered as an option for model and tool access.
Capabilities
- Creates and improves its own skills during use
- Persistent memory with searchable past sessions
- One gateway for Telegram, Discord, Slack, WhatsApp, Signal and the command line
- Built-in scheduler for unattended tasks
- Works with the model provider you choose, including your own endpoint
- Seven terminal backends, from local to serverless sandboxes
When to use it
- You want an agent that keeps context and builds up reusable skills over time
- You want to message your agent from a chat app while it works on another machine
- You want to switch model providers without changing code or being locked to one
