AI agents (Hermes)

AI agents that get work done — not just chat

An AI agent is more than a chatbot: it plans steps, uses tools, remembers context and works in your channels. We deploy agents built on Nous Research’s Hermes Agent and n8n.

An AI agent is software where a language model doesn’t just answer questions but decides which steps to take to reach a goal: it searches documents, calls tools (CRM, calendar, browser, APIs), remembers decisions and reports back. We deploy agents built on Hermes Agent — an open-source (MIT-licensed) agent from Nous Research with cross-session memory, a skills system and messaging integrations — as well as agents built in n8n.

Agent in action

One instruction. The agent does the rest.

The agent gets a goal, plans the steps, uses tools (CRM, email, calendar), saves proven approaches as skills and schedules recurring jobs. You approve what matters.

hermes — telegram · slack
  1. YouPrepare a Monday report: new leads from last week that haven’t had a reply yet, and suggest who to contact first.
  2. planPlan: 1) fetch last 7 days of leads from the CRM, 2) check email history, 3) score potential, 4) prepare the report and draft replies.
  3. toolcrm.search(created > -7d, status = "new") → 14 records
  4. toolgmail.threads(to: leads) → 5 without a reply
  5. toolskill: lead-scoring (saved skill) → scores 92, 81, 64, 40, 22
  6. HermesDone. 5 leads are waiting. Start with: the fashion store (92, wants to launch in December) and the clinic (81, needs a site + local SEO). Draft emails are in your inbox. I scheduled the report for Monday 8:00 — you’ll get it on Slack.
  7. toolcron.add("0 8 * * 1", report.weekly_leads) ✓
Sample agent session (illustration on test data).

Agent in n8n

Or an agent wired into your workflows

Drag the nodes · click “Run”

Execution log

    Support agent: WhatsApp message → AI agent with tools → reply or human hand-off

    Why us

    What you get

    Works in your channels

    The agent works in Slack, Telegram, WhatsApp or Discord — where your team already is.

    Memory and skills

    Hermes Agent keeps context across conversations and saves proven ways of solving tasks as reusable “skills”.

    Scheduled tasks

    Morning sales report, ticket review, competitor monitoring — the agent runs them on schedule and delivers the result.

    Control and safety

    Clear permissions, sandboxed environments, approval for risky actions and logs of everything. The agent only does what you allow.

    Scope

    What exactly we do

    Tools & technology

    • Hermes Agent
    • n8n
    • OpenRouter
    • Anthropic
    • OpenAI
    • MCP
    • Docker
    • Telegram
    • Workshop: which tasks to hand to an agent
    • Hermes Agent deployment on a server (Docker/VPS) or an n8n agent
    • Model setup (OpenAI, Anthropic, OpenRouter, local models)
    • Agent tools: CRM, calendar, email, knowledge bases, APIs
    • Messaging integrations (Slack, Telegram, WhatsApp, Discord)
    • Safety rules, permissions and logging
    • Testing on real scenarios and ongoing tuning

    FAQ

    Frequently asked questions

    How is an AI agent different from a chatbot?

    A chatbot answers questions from a script or knowledge base. An agent gets a goal and plans the steps itself: it uses tools, checks results, corrects course and finishes the job — e.g. books a meeting, prepares a report or updates the CRM.

    What is Hermes Agent?

    Hermes Agent is an open-source AI agent developed by Nous Research under the MIT licence. It has built-in cross-session memory, a skills system, task scheduling, many tools and connects to messaging apps. It works with multiple model providers.

    Can the AI agent run on our own server?

    Yes. Hermes Agent and n8n can run on your own server, using an external API or a local model depending on your data requirements.

    What do businesses use AI agents for?

    Most often for first-line enquiries, research (market, competitors, leads), reporting, inbox triage, sales support and admin tasks.

    How do you keep an agent safe?

    We keep permissions minimal, run the agent in a sandbox, require approval for risky actions (like emailing customers) and log every operation.

    Ready for lift-off?

    Tell us about your project. We reply within one business day — with ideas, not a boilerplate offer.