What is Hermes Agent?
Hermes Agent is Nous Research's self-hosted agent. You install it on your machine or a small VPS, point it at a model provider such as Nous Portal, OpenRouter or OpenAI, and talk to it from a terminal UI, Hermes Desktop, or messaging apps including Telegram, Discord and Slack. It stores memory and skills across sessions. The agent software is MIT-licensed; model and hosted-tool use are billed through Nous Portal or your own keys.
What you use Hermes Agent for
Hermes Agent is built by Nous Research as an autonomous agent you host yourself. The docs describe it as not tied to a single IDE: you install it, choose a model provider, and keep chatting across sessions while it builds memory and reusable skills. The README lists a full TUI, a messaging gateway covering Telegram, Discord, Slack, WhatsApp, Signal and CLI, scheduled automations in natural language, and subagents that run isolated workstreams.
You pick models with `hermes model` and can use Nous Portal, OpenRouter, OpenAI, your own endpoint and other listed providers. Terminal work can run locally or on Docker, SSH, Modal, Daytona, Vercel Sandbox and plugin backends. Hermes Desktop is offered for macOS, Windows and Linux alongside the CLI installer.
Plans and what they include
The agent itself is open source under the MIT license on GitHub (NousResearch/hermes-agent). Model and hosted-tool costs depend on where you send traffic. The Hermes homepage pricing table for Nous Portal lists Free at $0 with pay-as-you-go and free models only; Plus at $20/month; Super at $100/month; Ultra at $200/month. It says all paid tiers include usage in Hermes Agent, access to 200+ models, built-in tool use and one cloud agent, with cloud RAM/disk rising from 4 GB / 25 GB on Plus to 8 GB / 100 GB on Ultra. Hosted tools such as web search, images, speech and cloud browser are marked available on paid tiers. Current plans are confirmed in Nous Portal. Using your own API keys with OpenRouter or OpenAI follows those providers' prices instead.
A first workflow to try
On Linux, macOS or WSL2, run the install script from the docs: `curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash`. On native Windows use the PowerShell installer from the same docs. Then follow the quickstart: choose a provider (the docs suggest `hermes setup --portal` for one OAuth covering a model plus Tool Gateway tools), start a chat, and verify a session. Ask the agent to create a small skill after a multi-step task and check that it still recalls it in a new session. AppFor.ai has not run this install; the steps follow Nous Research's documentation.
What stands out
- Terminal UI with streaming tool output, slash commands and conversation history
- Messaging gateway: Telegram, Discord, Slack, WhatsApp, Signal, email and CLI
- Memory, skill creation and cross-session search
- Cron-style scheduled jobs and isolated subagents
- Terminal backends: local, Docker, SSH, Modal, Daytona and others
Before you choose
- You host the agent; security of keys, messaging bots and sandboxes is your responsibility.
- Nous Portal prices and cloud sizes can change; check the Portal for the live table.
- Vendor claims about a unique learning loop are marketing language, not an independent benchmark.
- Example workflows come from Hermes docs and README, not AppFor.ai tests.
Go to the source
Sources & further reading
Based on the official documentation below. Features and access can change; check the provider before making a decision.
Was this page useful?
Help us make our coverage more useful. This is feedback on the page, not a rating of the tool.
Go a little further