Hermes · Managed by AI HUB

Hermes,
fully managed.

Hermes is an open-source AI agent that handles conversations across WhatsApp, Telegram, Slack, Discord and Signal. We run it on our infrastructure, so your team can start using the agent without a server, a developer, or any setup work.

Get started Talk to sales
WhatsApp Telegram + Discord, Slack, Signal
9:41 ● ● ●
H
Hermes
Online
Morning! What's on the agenda today?9:01 AM
3 meetings: Marketing standup at 10, design review at 2, dinner with Sarah at 7. I also drafted your weekly report — want me to send it?9:01 AM
Yes please. And remind me to call mom this evening.9:02 AM
Sent the report to your team. Reminder set for 6:30 PM — "Call mom" 💚9:02 AM
Message Hermes…
Why managed Hermes

What we handle for you.

Installing Hermes from scratch needs a server, Docker, Python and your own LLM gateway. We run all of that on our infrastructure so your team doesn't have to.

Infrastructure, managed

Your Hermes lives in a private container on our servers. We handle the install, updates, restarts and monitoring. Your team uses the agent through a web dashboard and never needs to log into a terminal.

Isolated and encrypted

Each customer runs in their own container. Your messenger credentials, agent memory and conversation history never share space with another customer's data. API keys and tokens are encrypted at rest.

One key, every model

Your Hermes can use Claude, GPT, Gemini, DeepSeek and Llama through a single AI HUB credential. If you'd rather bring your own keys, OpenRouter, any OpenAI-compatible endpoint, or a local vLLM work too.

What you're deploying

What Hermes does for your team.

Hermes remembers what you tell it, learns new skills as it works, and ships with a built-in toolkit. Tasks you teach it once become reusable for next time.

40+ built-in skills

Web browsing, vision, image generation, text-to-speech, scheduled cron jobs, parallel sub-agents, file handling and an agentic mailbox. All bundled and available on day one.

Persistent memory across channels

Hermes keeps context between sessions and across messengers. A Telegram conversation in the morning continues on WhatsApp at night without losing track of what was said.

Skills it writes itself

When Hermes learns a new workflow it records the steps as a reusable skill document (using the agentskills.io standard). Your team's processes become the agent's playbook.

Personal or team

Choose how your agent works.

Hermes can run as a private assistant for one person or as a shared agent for a whole team. Pick the mode that matches how your group already works.

Personal assistant

A private Hermes for one person — connected to your own messenger accounts, with memory and skills scoped to you. Good for founders, freelancers, and individuals who want a single agent across their personal channels.

  • Your own WhatsApp, Telegram or Discord accounts
  • Memory and skills tied to you alone
  • Billed per agent, per month
Coming soon

Team assistant

One Hermes shared across the whole team — with shared memory and skills, and separate conversation contexts per sender. Good for ops, support and marketing teams who want institutional knowledge in one place.

  • One agent for the whole organisation
  • Shared memory and skills across members
  • Per-sender conversation isolation built in
Available everywhere

One agent,
every messenger.

Same memory, same skills, same agent — wherever your team and customers already talk.

MIT Open source · Read the docs

WhatsApp

LIVE

Voice notes, image attachments and group chats — your Hermes agent handles them all from the world's most-used messenger.

2.8B+ users worldwide · Business API supported

Telegram

LIVE

Bot API with inline commands, file uploads and channel posting. A fit for power users and ops teams.

Discord · Slack · Signal

LIVE

One Hermes follows you from team channel to private chat — one shared memory.

Deployed for you

Hermes, deployed for you.

Start in five minutes — no server, no setup, no surprise bills. Built on the open-source Hermes Agent and managed end-to-end by AI HUB.

  • 30-day money-back guarantee
  • MIT-licensed core
  • No card to start
Get started Talk to sales