Nous Hermes and Hermes Agent
Nous Research is behind two things that are often confused: the Hermes family of open-weight models and Hermes Agent, a self-improving open-source agent that works with any model.
- Project
- Hermes Agent
- Publisher
- Nous Research
- Repository
- github.com/nousresearch/hermes-agent
- Documentation
- hermes-agent.nousresearch.com/docs/
- Languages
- Python
- Category
- AI agent frameworks
Nous Hermes: the model family
Under the Hermes name, Nous Research publishes open-weight language models available on Hugging Face – including the Hermes 3 and Hermes 4 generations in different sizes. Open-weight models can be self-hosted, which opens up new answers to questions of data sovereignty, cost and independence from individual providers.
Hermes Agent: the self-improving agent
Hermes Agent is an open-source agent by Nous Research with a built-in learning loop: after complex tasks it creates skills on its own, improves them in use, maintains memory and searches past conversations. The model is freely selectable – via Nous Portal, OpenRouter, OpenAI, your own endpoint and other providers.
- Channels
- Telegram, Discord, Slack, WhatsApp, Signal and CLI through a single gateway process.
- Scheduled automations
- A built-in scheduler for recurring tasks such as reports or audits.
- Delegation
- Isolated sub-agents for parallel workstreams.
- Execution environments
- Local, Docker, SSH and various cloud sandboxes.
What Hermes Agent is suited for
- Personal and team assistants reachable via messengers
- Recurring tasks that should run unattended on a schedule
- Scenarios in which the model should be freely selectable or self-hosted
- Experiments with agents that build their own capabilities from experience
Using Hermes effectively in an AI-first business model
An agent that creates new skills for itself and stores knowledge permanently changes the question of control. In business use, it should be defined which skills may be adopted, what the agent may remember and who reviews and corrects both.
Unattended tasks need clear limits: approved channels and tools, defined recipients and an escalation path when something unexpected happens. Open models can strengthen data sovereignty – but responsibility for operation then lies entirely with the company or its operator.
From tool to outcome
In AI-first coaching you clarify which outcomes AI agents should be accountable for in your go-to-market and within which framework. REVAGO then designs, builds and runs your system. You pay for results – not for activity.
Frequently asked questions
What is Nous Hermes?
Nous Hermes refers to the Hermes language models by Nous Research, published with open weights, including the Hermes 3 and Hermes 4 generations. It is distinct from Hermes Agent, an open-source agent by the same team that works with any model.
What is Hermes Agent?
Hermes Agent is a self-improving open-source agent by Nous Research. It creates and improves skills from experience, maintains memory, works via Telegram, Discord, Slack, WhatsApp, Signal and CLI, supports scheduled tasks and sub-agents and can run with different model providers.
Does Hermes Agent have to run on a Hermes model?
No. Hermes Agent is model-agnostic. The model can be selected via Nous Portal, OpenRouter, OpenAI, your own endpoint and other providers and switched without code changes.
Further reading in the wiki
A neutral overview, not a recommendation of individual providers. REVAGO is not affiliated with the listed projects. Descriptions are based on the projects’ public information; features and licences may change.