What if I want to switch it off, or something breaks?

You can pause an agent, revoke a connection, or remove the app from Slack or Teams — all from your side, in seconds. And if something breaks, your subscription covers it: we monitor, maintain and update your platform and keep your agents healthy. No tickets, no portals, no chasing.

We hatch AI agents for your business.

Q: What is Hermes?

Hermes is the AI platform we install in your business — the nest your agents hatch into. It comes with hundreds of pre-installed skills, connects to the tools you already use, and learns how your business sounds — so the work it produces sounds like you, not like a chatbot.

Q: Is my data safe? Do you need my passwords?

We never ask for passwords — we connect through each app's official 'Sign in with…' approval screen, and you can revoke any connection yourself at any time. Hermes is installed for your business alone — on its own Linode server, with your files in your own Google Cloud storage bucket — each agent gets access only to the tools its job needs, anything customer-facing waits for a person to approve it, and we don't use your data to train AI models.

Pick a job you're sick of. We hatch an agent that owns it — and it works in Slack or Teams like a colleague. One agent, job done.

Book a call → See your agent → Watch one work ↓

Your agents plug into the tools you already run No rip-and-replace

Watch one actually work.

One message in, finished work out — in the app your team already has open.

Same agents · same memory · switch any time

# invoice-generator

Invoice Generator — in this channel

Model

Message #invoice-generator

Each channel is its own agent — @mention Hatchy like a teammate and finished work comes back. Swap the model with one click; the agent keeps its memory, skills and voice.

Update copy, hours and pages
Publish new blog posts
Swap images and banners
Build a staging preview first
Go live on your approval

Back

Custom

Need something else?

We'll hatch a custom agent for your exact job — quoted up front. Or build your own in plain English, free, forever. No code.

Ask us →

The Difference.

A model answers questions. An agent gets the job done. Here's the difference, in plain English.

What we're comparing	Your hatched agent	ChatGPT/Claude
In your Slack & Teams?	✓Yes — it works where your team already talks	–No — its own separate app
Whose brand?	✓Yours — your agent, your name, your voice	–Theirs — everyone knows that logo
One agent or a team?	✓A team — one specialist agent per job	–One chat window for everything
Who gets to talk to it?	✓You decide — per team, per person, per agent	–A seat is a seat — everyone gets the same
The harness	✓Skills, memory & approvals around the model — that's what makes it an agent	–None — just the raw model
Open source?	✓Yes — you can see exactly how it works	–No — you can't look inside

Your agents run on the same models that power ChatGPT and Claude. The difference is everything wrapped around them — your tools, your voice, your approvals.

Security.

It reads invoices and sends email as your company. Here's how we keep that safe — no acronyms.

Hatchy resting a hand on its own little storage cabinet

Your own server, your own storage

Hermes isn't a shared app you log into next to a thousand other companies. It runs on its own Linode server, set up for your business alone, and your files live in your business's own Google Cloud storage bucket. Everything is encrypted on the way there and while it sits there, nobody else's data shares the space — and if you ever leave, it's handed back to you and wiped from our side.

One job, one set of keys

The invoice agent can reach your accounting software — and nothing else. Like a hotel keycard: it opens your room, not every door in the building. And we never ask for passwords: every connection goes through the app's official "Sign in with…" screen, so you can revoke it yourself.

Hatchy holding an envelope, waiting for approval

Nothing sent without you

Agents draft and prepare on their own, but anything that reaches a customer or moves money waits for a human to click approve.

Your data isn't training material

Your emails and documents do your job and nothing else — we don't use them to train AI models. AI requests run pay-per-use through paid business APIs, not through free consumer chatbots.

You hold the plug

Pause an agent, revoke a connection, remove the app — from your side, in seconds.

Full detail in our Privacy Policy and Terms. Harder questions? Ask us directly.

Questions people ask.

What exactly is Hermes?+

The AI platform we install in your business — the nest your agents hatch into. It comes with hundreds of pre-installed skills, connects to the tools you already use, and learns how your business sounds. So the work it produces sounds like you, not like a chatbot.

What's an AI agent?+

A teammate in software form. Each agent has one job — chasing leads, writing documents, processing invoices, updating your website. You talk to it in plain English in Slack or Microsoft Teams, and it does the work. Because each agent only does its one job, it stays focused and never gets confused.

How much does it cost?+

Three tiers: Basic is A$10,000 and then A$499/month. Standard is A$25,000 with our agent marketplace to pick and hatch from, then A$499/month. Premium starts from A$50,000 for bespoke enterprise agents. Prefer not to own it? Rent the platform for A$949/month with no install fee — agents not included. AI usage is billed at cost through OpenRouter: you pay for what your agents actually use, not a per-seat licence.

Why not just use ChatGPT or Claude?+

Three reasons. Cost: those are fixed per-seat subscriptions you pay whether you use them or not — Hermes is pay-per-use, so quiet weeks cost almost nothing. Voice: generic chatbots sound generic — your agents learn your business, so everything sounds like you. Flexibility: when a better or cheaper model comes out, you swap it in with one click. You're never locked in. The comparison table has the whole argument.

Do I need to be technical?+

No. If you can send a message, you can use your agents. We install everything, connect your tools, and hand it over working. You can also build your own agents in plain English — no code — or ask us to hatch new ones for you.

Is my data safe? Do you need my passwords?+

Never a password — we connect through each app's official "Sign in with…" approval screen, and you can revoke any connection yourself, at any time, from your own admin settings. Hermes is installed for your business alone — on its own Linode server, with your files in your own Google Cloud storage bucket — each agent gets access only to the tools its job needs, anything customer-facing waits for a person to approve it, and we don't use your data to train AI models. More in the security section.

What if I want out, or something breaks?+

Pause an agent, revoke a connection, or remove the app from Slack or Teams — all from your side, in seconds, and your data comes with you. If something breaks, your subscription covers it: we monitor, maintain and update your platform and keep your agents healthy. No tickets, no portals, no chasing.

Ready to hatch your first agent?

Thirty minutes, same week. Bring the job you're sick of doing — we'll tell you exactly what hatching it looks like.

Book a call → See your agent →