1 · Why we're building this
Most AI products were designed for an office: shared accounts, shared context, shared output, shared everything. That works at work. It does not work in a home. Your calendar, your finances, your conversations with the people you love — those don't want to share a room with thousands of strangers.
Ermes is the version where each person gets a private agent that lives on a server in their name. It uses the messengers they already use. It remembers what matters. It asks before it does anything irreversible. It is, on purpose, dull and reliable rather than loud and clever.
2 · Who we are
Built in Silicon Valley by a small team. We'll add names here as the team grows. Our investors and advisors are listed when they ask to be; quietly otherwise.
3 · How we work
- Customer's machine, customer's risk. Ermes touches infrastructure that belongs to other people. We default to conservative behaviour with their VPSs, secrets, and data.
- Boring tech that ships. We pick mature tools over novel ones unless there's a real reason.
- Plain English, including in the legal pages. If you can't understand a sentence on this site, that's our fault.
- One email when something matters. No drip campaigns, no engagement loops.
4 · Where we are
Ermes is in private beta in spring 2026. Invites go out in small batches. Get on the list if you want one.
5 · Get in touch
Press, partnerships, “is this a fit for my company”, or just want to say hi: write to [email protected] or use the contact page. A human reads it.
6 · The company
AgentHut Labs Inc. is a Delaware C-corp, built in Silicon Valley.