me
Persistent memory
Long-term context that survives sessions, devices, and restarts. The agent remembers what you told it last week.
Now in private beta · Hosted Hermes Agent
Your self-improving
AI coworker, live in minutes.
A managed Hermes Agent with persistent memory, reusable skills, scheduled jobs, and chat-app access. No VPS. No Docker maze. No gateway plumbing.
No spam. One email if and when it ships.
Built on open-source Hermes Agent
Any frontier model
Telegram · Slack · Web
hermes://you · live● online
$ hermes setup --hosted
▸ provisioning private workspace…
▸ wiring telegram + slack gateways…
▸ loading memory · skills · cron · mcp…
✓ agent online — reachable from anywhere
you draft my Monday brief and DM it at 7am
hermes scheduled. learning your tone from past briefs
latency
~280ms p50
memory
persistent · yours alone
skills learned · last 7 dayslive
- skillsummarize_threadv3 · refined tone
- cronmorning_briefscheduled · 7:00 IST
- toolcompetitor_watchadded 2 sources
TelegramSlackWebOpenAIAnthropicOpenRouterMCPCron
Daily operator briefingsTelegram-first research agentScheduled audits & reportsInbox & calendar workflowsFounder ops assistantPersistent coding copilotTeam memory & handoffsPersonal automation OSCompetitive intel monitoringCustomer-conversation digestsDaily operator briefingsTelegram-first research agentScheduled audits & reportsInbox & calendar workflowsFounder ops assistantPersistent coding copilotTeam memory & handoffsPersonal automation OSCompetitive intel monitoringCustomer-conversation digests
Capabilities
Everything Hermes does, none of the hosting.
sk
Self-improving skills
Reusable, versioned capabilities the agent learns and refines from your usage — not throwaway prompts.
cr
Scheduled jobs
Cron-style autonomy. Daily briefs, weekly audits, hourly checks — running while you sleep.
ch
Chat-app native
Reachable from Telegram, Slack, and the web. Your agent lives where you already do.
to
Tools & MCP
Browser, search, code, files, and any MCP server. Real actions, not just answers.
mo
Any frontier model
Pick GPT, Claude, Gemini, or open models per workflow. Use the included pool or bring your own API key — no lock-in.
How it works
From zero to agent online in three steps.
The setup that takes a weekend self-hosted takes a few minutes here. We do the boring part.
- 01
Claim your instance
We provision a private workspace, wire the gateways, and load memory + skills + cron.
- 02
Connect your channels
Telegram, Slack, web — or all of them. Add MCP tools and the model you prefer.
- 03
Send the first message
Ask, schedule, automate. Your agent gets sharper every week — without any infra work from you.
The wedge
Hermes is powerful. Running it isn't.
Self-hosted Hermes
- Provision a VPS, configure Docker
- Wire gateways and providers
- Babysit uptime, backups, updates
- Debug at 2am when it dies
HermesCloud
What we do- Private instance, already running
- Channels and tools pre-wired
- Updates and backups handled
- Open the chat. Done.
Early access
Skip the setup. Get the agent.
Join the waitlist for early access and founding-user pricing.
No spam. One email if and when it ships.
FAQ
Honest answers.
What is Hermes Agent?
Hermes Agent is an open-source, self-improving AI agent from Nous Research with persistent memory, reusable skills, scheduled jobs, tools, and chat-app access.
How is this different from running Hermes myself?
No VPS provisioning, Docker debugging, gateway configuration, provider wiring, updates, or uptime work. You get the agent — not another infrastructure project.
Is this affiliated with Nous Research?
No. HermesCloud is an independent managed hosting layer built around the open-source Hermes Agent ecosystem.
What happens if there is real demand?
We invite early users, learn which workflows they actually want, and ship the smallest hosted Hermes service worth paying for.
When does it launch?
Soon. Join the waitlist for early access and founding-user pricing — you'll hear from us first.