Powered by Flux Decentralized Cloud

Hermes On the Decentralized Cloud

Your Self-Improving AI Agent. Fully Yours.

Deploy Hermes Agent, the self-improving AI agent by Nous Research, on the Flux decentralized cloud. Password-protected dashboard, built-in Tailscale to reach your private network, and bring your own AI provider key — OpenRouter, OpenAI, Anthropic and more. Your agent, your keys, your infrastructure.

First week free for new users
BYO
AI Key
Privacy
First
Instant
Deploy

What is Hermes Agent?

A self-improving AI agent built by Nous Research that plans and runs multi-step tasks on its own. Bring your own AI provider key, reach your private network over built-in Tailscale, and drive it all from a secure, password-protected dashboard you fully control.

Self-Improving Agent

Self-Improving Agent

Hermes by Nous Research learns from its own runs, refining its approach over time to handle complex multi-step tasks autonomously

Bring Your Own AI Key

Bring Your Own AI Key

Plug in OpenRouter, OpenAI, Anthropic (Claude), or any provider — your keys, your models, your billing, never touched by us

Built-in Tailscale

Built-in Tailscale

A Tailscale mesh ships inside the container so your agent can securely reach your private network and self-hosted services

Password-Protected Dashboard

Password-Protected Dashboard

Drive your agent from a secure, password-protected web dashboard — chat, inspect runs, and manage tools from anywhere

Privacy-First

Privacy-First

Self-hosted on decentralized infrastructure — your conversations, keys, and agent memory stay entirely under your control

Open API Server

Open API Server

A built-in HTTP API lets you wire Hermes into your own apps, scripts, and automations with full programmatic control

Why Choose Us

Experience the power of decentralized cloud infrastructure built on Web3 technology

Instant Deployment

Instant Deployment

Launch your Hermes agent in under 30 seconds with automated deployment

DDoS Protection

DDoS Protection

Enterprise-grade DDoS protection keeps your agent online during attacks

Global Server Network

Global Server Network

Choose from 50+ countries worldwide to run your agent close to your data

24/7 Monitoring

24/7 Monitoring

Real-time performance monitoring with automatic alerts

Pay-As-You-Go Pricing

Pay-As-You-Go Pricing

Only pay for resources you use with transparent, flexible pricing

99.9% Uptime

99.9% Uptime

Distributed network architecture ensures maximum availability

Cost Effective

Cost Effective

Run a dedicated AI agent for a fraction of managed-cloud pricing

Decentralized Infrastructure

Decentralized Infrastructure

No single point of failure across thousands of independent nodes

Simple, Transparent Pricing

Choose the perfect plan for your needs. Deploy in minutes.

🎁 First week free for new users*
40% OFF

Hermes Agent

$6.70
$4.02
per month
RAM4 GB
Storage
20 GBSSD/NVMe
CPU2 vCores
BandwidthUnlimited
40% OFF
Recommended

Hermes Agent Pro

$12.48
$7.49
per month
RAM8 GB
Storage
80 GBSSD/NVMe
CPU4 vCores
BandwidthUnlimited
First Week Free*
30-day money-back*
No setup fees
Cancel anytime

*The first week is free for customers new to Flux Cloud: one free week per account, not per app or per server. It takes no card and ends on its own after 7 days unless you choose to keep the server. If you've deployed any app on Flux before, standard pricing applies. Our 30-day money-back guarantee covers your first paid month, whether or not you took the free week first.

Looking for other apps?Browse the Flux Marketplace

Why self-host on the Flux decentralized cloud

Privacy, data ownership, bring-your-own-key and no lock-in — explained.

Learn more
Decentralized Infrastructure

Global Server Network

Your server runs on a decentralized network of 6,390 servers across 58 countries

0
Active Servers
0
Countries
99.9%
Uptime %
GermanyGermany1,430
United StatesUnited States1,014
FinlandFinland879
FranceFrance636
DenmarkDenmark547
CanadaCanada350
RussiaRussia195
PolandPoland140
United KingdomUnited Kingdom109
SpainSpain92
NorwayNorway76
UkraineUkraine69
SwitzerlandSwitzerland56
LatviaLatvia52
JapanJapan48
SwedenSweden40
United Arab EmiratesUnited Arab Emirates40
RomaniaRomania35
ChileChile33
SingaporeSingapore32
AustraliaAustralia28
ItalyItaly25
South AfricaSouth Africa23
SlovakiaSlovakia22
BulgariaBulgaria21
PortugalPortugal20
ArgentinaArgentina16
HungaryHungary16
BelgiumBelgium15
Hong KongHong Kong12
MalaysiaMalaysia9
EstoniaEstonia8
TaiwanTaiwan8
LithuaniaLithuania6
IsraelIsrael6
IrelandIreland4
VietnamVietnam4
BrazilBrazil4
MoldovaMoldova4
GreeceGreece3
MexicoMexico3
LuxembourgLuxembourg3
GermanyGermany1,430
United StatesUnited States1,014
FinlandFinland879
FranceFrance636
DenmarkDenmark547
CanadaCanada350
RussiaRussia195
PolandPoland140
United KingdomUnited Kingdom109
SpainSpain92
NorwayNorway76
UkraineUkraine69
SwitzerlandSwitzerland56
LatviaLatvia52
JapanJapan48
SwedenSweden40
United Arab EmiratesUnited Arab Emirates40
RomaniaRomania35
ChileChile33
SingaporeSingapore32
AustraliaAustralia28
ItalyItaly25
South AfricaSouth Africa23
SlovakiaSlovakia22
BulgariaBulgaria21
PortugalPortugal20
ArgentinaArgentina16
HungaryHungary16
BelgiumBelgium15
Hong KongHong Kong12
MalaysiaMalaysia9
EstoniaEstonia8
TaiwanTaiwan8
LithuaniaLithuania6
IsraelIsrael6
IrelandIreland4
VietnamVietnam4
BrazilBrazil4
MoldovaMoldova4
GreeceGreece3
MexicoMexico3
LuxembourgLuxembourg3

What is a self-improving AI agent?

A self-improving AI agent doesn’t just answer a prompt and stop — it pursues a goal. It plans a multi-step task, calls tools and APIs to carry each step out, checks its own results, and revises its approach when something fails. Over repeated runs it learns from its own work, so the jobs you give it get faster and more reliable. Hermes Agent by Nous Research brings that loop together with bring-your-own-key model access and built-in private-network reach, all on infrastructure you control.

How it works

1

Pick a plan & deploy

Choose Hermes Agent or Pro and deploy to the Flux decentralized cloud in about 30 seconds — no server setup.

2

Add your AI key

Paste your own OpenRouter, OpenAI, or Anthropic (Claude) key. The agent calls your provider directly with your account.

3

Drive it your way

Give it tasks from the password-protected dashboard, or call the built-in HTTP API from your own apps and scripts.

Why self-host instead of a managed agent SaaS?

With a managed agent SaaS your prompts, data, and keys flow through someone else’s backend, and you’re locked into their models, limits, and pricing. Self-hosting Hermes flips that trade: you get a dedicated, private instance, direct control of your model spend, and the freedom to move or shut it down at will.

Your data stays yours

Conversations, memory, and keys live on your instance — never routed through a shared SaaS backend.

Bring-your-own-key economics

Pay your AI provider directly at their rates. No per-token SaaS markup on every request.

Reach your private network

Built-in Tailscale lets the agent act on internal APIs and databases without exposing them publicly.

No vendor lock-in

Your instance runs on independent nodes. Switch models, migrate, or cancel whenever you like.

Frequently Asked Questions

Everything you need to know about our hosting service

Hermes Agent is a self-improving AI agent built by Nous Research. It plans and executes multi-step tasks autonomously, learns from its own runs, and can connect to your tools and private network — all driven from a secure web dashboard that you fully control.
Yes — Hermes is bring-your-own-key. Plug in an OpenRouter, OpenAI, Anthropic (Claude), or other provider key and the agent uses your account and models directly. Your keys live on your instance and your model usage is billed by your provider, never by us.
Each Hermes deployment ships with Tailscale inside the container. Join it to your tailnet and the agent can securely reach private services — internal APIs, databases, home-lab machines — without exposing them to the public internet.
The dashboard is password-protected and your instance is self-hosted on decentralized Flux infrastructure. Your conversations, provider keys, and agent memory stay on your own deployment — nothing routes through a shared SaaS backend.
We offer Hermes Agent (2 vCPU / 4 GB RAM / 20 GB) and Hermes Agent Pro (4 vCPU / 8 GB RAM / 80 GB). Pick Pro for heavier workloads, larger context handling, and more concurrent tool use. You can upgrade anytime.
Flux is a decentralized cloud infrastructure running on thousands of independent nodes across 50+ countries worldwide, providing redundancy and eliminating single points of failure.
We use transparent pay-as-you-go pricing with your first week free. Plans start at competitive rates with no hidden fees or long-term contracts. You only pay for the compute your agent uses.
Yes. Hermes is model-agnostic and bring-your-own-key. Paste an OpenRouter, OpenAI, or Anthropic (Claude) API key into your dashboard and the agent calls that provider directly with your account. Because keys stay on your instance and requests go straight to the provider, you keep full control of which models you use and what they cost — switch providers any time without redeploying.
A self-improving agent like Hermes plans a task, breaks it into steps, calls tools or APIs, checks its own results, and adjusts — then reuses what it learned on the next run. In practice that means researching and summarizing, writing and refactoring code, monitoring systems, orchestrating multi-step automations, and driving your own tools through its API, all without you scripting every step by hand.
With a managed SaaS your prompts, data, and keys flow through someone else's backend and you're locked into their models, limits, and pricing. Self-hosting Hermes on Flux gives you a dedicated instance you control: your own AI provider key, your own memory and conversation history, no per-seat SaaS markup, and the freedom to move or shut it down whenever you like. You trade a little setup for privacy, cost control, and no vendor lock-in.
Yes. Your Hermes instance is dedicated to you and password-protected, running on decentralized Flux nodes rather than a shared multi-tenant SaaS. Conversations, agent memory, and provider keys live on your deployment; model requests go directly from your instance to the AI provider you chose. Nothing is routed through, mined by, or retained on a central Hermes backend, because there isn't one.
Yes. Every deployment exposes a built-in HTTP API with an OpenAI-compatible surface, so you can drive the agent from your own apps, scripts, cron jobs, and automations. Send it tasks, stream responses, and manage sessions programmatically using your instance's API key — the same agent you use in the dashboard is available to your code.
Each Hermes deployment ships with Tailscale built in. Authenticate it to your tailnet and the agent can securely reach private machines and services — internal APIs, databases, home-lab hosts — over an encrypted mesh, without exposing anything to the public internet or opening firewall ports. It's the simplest way to let an AI agent act on your private infrastructure safely.
Hosting starts at $4.02/month pay-as-you-go with your first week free, and that runs your agent around the clock — no per-hour metering on the instance. Your only other cost is your own AI provider usage (OpenRouter, OpenAI, or Anthropic), billed by them at their rates. Because you bring your own key, you see and control that spend directly instead of paying a SaaS markup on every token.
Yes. Start on Hermes Agent (2 vCPU / 4 GB / 20 GB) and move up to Hermes Agent Pro (4 vCPU / 8 GB / 80 GB) whenever your workloads grow — heavier tool use, larger context, or more concurrent tasks. Pricing stays pay-as-you-go, so you only scale up when you need to.

Still have questions?

Join our Discord community →

Hermes Agent Guides

Everything you need to understand, deploy, and self-host Hermes Agent.

Support & help centerOpen a support ticket about deployments, billing, or managing your Hermes instance.