Now Live

Your AI agents,
always on.

HermesOS gets your Hermes agent running in under 5 minutes — with persistent memory, browser automation, and tool use. No Docker, no config files, no midnight debugging.

Free tier is live. Pro and Power tiers are available for serious workloads. Pay with card or $HermesOS.

Always free tier
BYO key — no markup
Built on Hermes Agent
Built on Hermes Agent (Nous Research)BYO key, zero markupFree tier always available

OpenClaw forgets. Hermes compounds.

Built by Nous Research, Hermes lives on a server and remembers everything — projects, preferences, lessons learned. Every session makes it sharper. Getting it self-hosted takes most people a weekend.

HermesOS takes that weekend down to 5 minutes.

What's included

Everything your agent needs. Nothing you don't.

Zero config. Full stack.

Browser automation, tool use, terminal, memory, and cron — pre-configured. No Docker, no StackOverflow at midnight.

Multi-agent from day one.

Unlimited agent profiles on a single instance. Researchers, operators, specialists — no extra cost per agent.

Your key. Zero markup.

OpenRouter, OpenAI, or Anthropic. Encrypted at rest, injected at deploy. We never see your AI spend.

Chat from anywhere.

Built-in dashboard with streaming. Connect Telegram, Discord, Slack, or WhatsApp — out of the box.

OpenClaw migration built in.

Your existing setup, prompts, and skills transfer over intact. No starting from scratch.

Stable. Recoverable. Always on.

Updates tested against your config before rollout. Auto-restart on failure. Daily backups — never more than 24 hours from clean restore.

How it works

Three steps. No terminal required.

1

Pick your tier

Start free, upgrade when you need more. Pay monthly with card, yearly for a discount, or hold $HermesOS to maintain access without a subscription.

2

Add your AI key

Paste your OpenRouter, OpenAI, or Anthropic key once. Encrypted, injected, done.

3

Deploy. Talk. Automate.

Your agent is live in minutes — chat, terminal, monitoring. Connect Telegram or Discord and it follows you everywhere.

Ready to move from reading to setup? Pick your tier and go straight into account creation.

Pick Your Tier
Use cases

What does a 24/7 agent with memory actually do?

Browse, code, manage files, call APIs, run cron tasks — autonomously. And it remembers what it learned last week.

DevOps & Monitoring

Reads logs, restarts failing services, pages you only when a human is actually needed. Remembers your stack.

Research & Competitive Intel

Give it a topic and a deadline. It browses, aggregates, and delivers a structured brief — then saves what it learned for next time.

Background Automation

Email triage, scheduling, API calls, spreadsheets — anything repetitive. Runs on cron while you sleep.

Customer Support Triage

Feed it your docs. It resolves common tickets, escalates the hard ones, and gets smarter with every conversation.

Seen the kind of work Hermes can take off your plate? Choose a plan and launch the workflow that matches your load.

See Plans & Launch
What's coming

This is just the beginning.

The hosting is the foundation. Coming soon:

Operator Packs

Pre-built agent templates for specific work — research, trading intel, content automation. Deploy in one click.

Marketplace

Build operator packs, ship them to the community, earn from usage. Settled in $HermesOS.

Agent Endpoints

Expose your agent as a callable API. Other agents pay you per request.

Hive Mind

Agents share what they learn. The whole network gets smarter together.

HermesOS is the infrastructure for an agent economy. Hosting is step one. Everything else builds on top.

Pricing

Simple plans. Serious compute.

Dedicated compute. Unlimited agent profiles. BYO key, zero markup.

Get Started

Free

Most users can launch without a card; higher-risk free-tier deploys may need card verification first.

$0

Always free; card only if risk checks require it

Compute
vCPU0.5
RAM1 GB
Active agents1
  • Persistent memory
  • All integrations included
  • Fair-use limits apply
Most PopularRecommended

Pro

For real work, not just experiments.

$9.99/mo

Subscribe monthly with card

Compute
vCPU2
RAM4 GB
Concurrent agents3
  • Unlimited agent profiles
  • Everything in Free
  • Priority over free tier
Three ways to access Pro
  • Monthly card
    $9.99/mo
  • Yearly
    $79/yr card · $49/yr in $HermesOS
  • Hold $HermesOS
    ~$99 (launch rate, first 30 days)
For Pros

Power

For serious workflows and multi-agent operations.

$19.99/mo

Subscribe monthly with card

Compute
vCPU4
RAM8 GB
Concurrent agentsUnlimited
  • Unlimited agent profiles
  • Everything in Pro
  • Burst CPU when capacity allows
Three ways to access Power
  • Monthly card
    $19.99/mo
  • Yearly
    $149/yr card · $99/yr in $HermesOS
  • Hold $HermesOS
    ~$199 (launch rate, first 30 days)

Save up to 40% paying with $HermesOS. Launch pricing for the first wave — rates may adjust as the platform matures.

Free tier — try before you upgrade · 48-hr refund on card payments
About $HermesOS

The access layer for the platform.

$HermesOS lets you pay subscriptions for a discount, hold to maintain your tier without paying monthly, or transact across the agent economy as we ship it.

The token isn't required to use HermesOS — the free tier stays available without it, with anti-abuse checks where needed. If you want to be in the ecosystem deeper, holding $HermesOS unlocks more flexibility on how you pay and access.

Common questions

Straight answers.

Is my API key safe?

Yes. Encrypted at rest, injected at deploy via environment variables. We never proxy or log your AI requests.

Will updates break my setup?

No. Every update tested against container configs before it ships. Daily backups mean you're never more than 24 hours from a clean restore.

How is this different from OpenClaw?

OpenClaw is an excellent open-source desktop framework. HermesOS is a fully managed, production-grade cloud environment. Hermes Agent has more stable memory, higher reliability, and updates that don't break your existing features.

What if my agent crashes?

It restarts automatically. Health, logs, and resource usage always visible in the dashboard.

Can I run multiple agents on one plan?

Yes — unlimited profiles per instance. Free runs 1 active agent, Pro runs 3, Power has no concurrent cap. Your compute pool is the only real limit.

Which AI providers are supported?

OpenRouter, OpenAI, and Anthropic. OpenRouter alone gives you hundreds of models on a single key.

Do I need to use $HermesOS to access the platform?

No. Free tier doesn't require it. Pro and Power can be paid with card. The token gives you discounts and a third payment path for people who want it.

What's coming next?

Operator packs (pre-built agent templates) ship in the next few weeks. Marketplace, agent endpoints, and Hive Mind follow. The roadmap is on hermesos.cloud/roadmap.

Start now

Ready to deploy?

Free tier is live with abuse safeguards. Pro and Power are available now.

Most free users can start without a card; higher-risk signups may need a card-on-file check.

Already have an account? Log in