Hermes Agent · expertise and hands-on support, based in France

Hermes Agent for business. The open source AI agent that works for you.

Hermes Agent is an open source AI agent released by Nous Research: it acts inside your tools, keeps a memory, learns reusable know-how and can run tasks on a schedule. At Kailli, it works on our own projects every day. We help you install it, set its guardrails and train your teams.

Book 20 minutes with us See what it does for us

Hermes Agent is an open source project by Nous Research. Kailli is not affiliated with Nous Research.

01In two sentences

What is Hermes Agent?

An assistant like ChatGPT tells you how to do it. Hermes Agent does it: it reads your files, runs commands, writes into your tools, then hands you the result.

  • A memory: it notes your preferences and what has been decided.
  • Skills: written know-how cards it reuses without being told twice.
  • Scheduled tasks: "every Monday, prepare the weekly summary."
  • Guardrails: dangerous commands wait for your approval.

It works with the model of your choice (Anthropic, OpenAI, OpenRouter or a local model) and you drive it from a terminal or a messaging app such as Discord, Slack or Telegram. For the basic idea, without jargon: an AI agent explained simply.

02In-house

What Hermes Agent already does at Kailli

We don't recommend a tool we don't use. Three real examples, shown without client names or private data.

The site you are reading

Kail.li was designed, written, checked and published with Hermes Agent: pages, automated quality checks, deployment prep. A human reviews and decides when it goes live.

The founder's second brain

A knowledge base kept up to date by the agent: context, decisions, meeting notes and next actions, all one question away.

SEO research before writing

Keyword and Google results research before every page. This Hermes Agent guide came out of it: demand was measured before the first line was written.

See our detailed use cases

03Use cases

Six ways to use Hermes Agent at work

An agent pays off when the task comes back often, follows known steps and can be checked. Six examples we set up or already use.

  • Build and maintain a website, with checks before every release.
  • Prepare sales follow-ups from the CRM, sent by a human.
  • Keep the team's knowledge base: decisions, meeting notes, actions.
  • Write and publish articles in the CMS, after review.
  • Prepare invoices and deposits in the accounting tool, approved before sending.
  • Monitor a server: audits, backups, alerts in your messaging app.

Each use case in detail, with its human safeguard

04Installation

Install Hermes Agent without taking risks

The installation itself takes a few minutes. What needs care is the framework: where it runs, what it sees, what it is allowed to do.

  • On a laptop, to try it out, or on a server in Docker, for a team.
  • A model chosen according to your data and your budget.
  • Secrets kept out of its reach, permissions limited to what is needed.
  • Human approval on anything that sends, pays or deletes.

Steps, skills, memory and security: how to install Hermes Agent

05Comparison

Hermes Agent or OpenClaw?

Two MIT-licensed open source agents you drive from your messaging apps. Hermes Agent bets on learning (memory and skills built through use); OpenClaw on a personal assistant present on all your devices.

The factual, dated comparison: Hermes vs OpenClaw

06For business

Why Hermes Agent for business

Plenty of tools promise agents. Hermes Agent ticks the boxes that matter when you install it for a team, not just for a demo.

Your data stays with you

It installs on your computer or your server. Memory, skills, history and settings are files you keep, back up and read. No subscription to an agent vendor.

The model of your choice, and the freedom to switch

Anthropic, OpenAI, Google, OpenRouter, Nous Portal or a local model: you choose based on data sensitivity and budget, and you switch without rebuilding everything.

Know-how that stays in the company

Every procedure it learns becomes a written skill a human can read. When someone leaves, the method stays; when the agent gets it wrong, you fix the card.

Adjustable security controls

Approval of dangerous commands, a list of authorized users, isolated execution in Docker or on another machine, secrets out of its reach. The mechanisms exist; we configure them with you.

Where your team already works

Discord, Slack, Microsoft Teams, Telegram, email: the team writes to it from their usual tools, with no new app to adopt.

Open source, MIT license

The code is public and auditable, with no lock-in to a single vendor. If you change tools tomorrow, your cards and rules remain usable.

The flip side: an agent that acts needs written rules and someone to enforce them. That is exactly the part we take on with you.

07How we help

Training, installation and security scoping

Kailli sells no tools and is not affiliated with Nous Research. We help you decide whether Hermes Agent is a fit, then get it running properly.

Training

Your teams learn to hand a task to an agent, write its skills and review its work. It is the core of our AI agents training, delivered by DAZZ SOLUTIONS, Qualiopi-certified for training activities (Qualiopi is France's national quality certification for training providers).

Guided installation

We install Hermes Agent with you, on your computer or your server, with a first real use case and written rules.

Security scoping

What the agent reads, what it does, who approves: three written answers before anything goes live, together with your IT department.

A project with Hermes Agent? Let's talk for 20 minutes, no commitment.

Book 20 minutes

Written by Rémi Carbonnefounder and CTO of Kailli

Frequently asked questions about Hermes Agent

What is Hermes Agent?

Hermes Agent is an open source AI agent released by Nous Research under the MIT license. It acts inside your tools, keeps a memory between sessions, creates skills from experience and can run scheduled tasks.

Is Hermes Agent free?

The software is free and open source. You pay for the AI model it uses (subscription or usage-based billing with the provider you choose) and, where applicable, the server that hosts it.

Is Hermes Agent suitable for a business?

Yes, if it is governed: limited accounts and permissions, secrets out of its reach, human approval of sensitive actions and written rules on data. Without that framework, we don't recommend it.

Which models can Hermes Agent use?

It works with models from Nous Portal, OpenRouter, OpenAI, Anthropic, Google or any OpenAI-compatible service, including a local model.

Is Kailli linked to Nous Research?

No. Kailli is not affiliated with Nous Research. We use Hermes Agent every day and support companies that want to adopt it.

Do you offer Hermes Agent training?

Hermes Agent is one of the tools in our AI agents training, in-house or in multi-company sessions. We can also install it with you and scope its security.

Twenty minutes. And you'll know whether we can help.

No sales pitch. You tell us where you stand with AI, we tell you how far you can go.

Book 20 minutes Prepare for the call