Agent infrastructure · v1

GiveyourteamaHermesagent—
with a permanent home.

alcione spins up a Hermes agent on a private VPS — wired to the model, memory, and voice you choose. One organization, one agent, one address. Built for teams that want their agents to be theirs.

Spins up in 90 seconds · No credit card to start

hermes agent · live

online

Aurora

Internal R&D copilot

MMODELLlama 3.3 70B Turbo

↯MEMORY8.4k memories

▶VOICERachel · narration

□VPSKVM 1 · São Paulo · BR-SP

Latency● 142 t/s · 1.2s

The stack

Six pieces. Wired up. Out of the box.

Don't stitch services. Don't write infra. alcione hands you a working agent on day one.

Private VPSINFRA

Your agent has its own address — a private server in the region you pick.

Hermes runtimeAGENT

Pre-installed. Tools, sessions, autonomous loops. Ready on first boot.

100+ modelsBRAIN

Pick any open-source LLM. Switch in one click. No vendor lock-in.

Persistent memoryRECALL

Semantic memory across conversations. Your agent remembers what matters.

Optional voiceSPEECH

A dozen studio-grade voices. Toggle on if your agent speaks back.

Team-firstCOLLAB

One organization, one agent. Invite teammates, set roles, share context.

How it works

From idea to deployed agent in 0 steps.

00ORG

Create your organization

POST /orgs{ "name": "Aurora ops" }

00CONFIG

Pick a model, memory tier & voice

Choose from 100+ open-source models. Pick a memory tier. Optionally give your agent a voice.

agent.configLlama 3.3 70B · team · Rachel

00DEPLOY

Hermes spins up on a fresh VPS

We provision the server, install the Hermes runtime, and wire everything in. Your agent is live in 90 seconds.

vps.boot✓ ready · 89.4s

Features

Everything an agent needs to feel real.

100+ MODELS

Switch your agent's brain.

Llama 3.3 70B. DeepSeek V3. Qwen 2.5. Mixtral. Gemma. Hot-swap in one click — your agent's memory and voice stay the same.

◆Llama 3.3 70B Instruct Turbo$1.50/M

◇DeepSeek V3$2.10/M

◇Qwen 2.5 72B Turbo$2.00/M

MEMORY

Your agent remembers.

Semantic memory across every conversation. Facts, preferences, tool approvals — automatically.

FACTUser's timezone is BRT (UTC-3)

PREFERENCEAlways use SI units in numbers

CONTEXTMigrating MongoDB → Postgres

VOICE

Optional — pick a voice.

A dozen studio-grade voices, English & multilingual. Toggle on if your agent speaks; off if it's silent.

+ 12 voices

8 REGIONS

Pick your address.

BR-SPUS-TXUS-AZDE-FRGB-LDNL-ASSGIN-MH

TEAM

Many hands, one agent.

Invite teammates to your org with roles. They share the agent.

YOUR VPS

SSH in. It's yours.

$ ssh user@aurora.alcione.ai

Use cases

Built for teams that need their own agent.

FOR STARTUPS

An internal copilot that knows your stack.

Onboards new engineers
Reads your codebase
Remembers every decision

FOR AGENCIES

One agent per client — full data isolation.

Their data, their VPS
Per-client model & memory
White-labeled voice

FOR INDIE DEVS

Your personal AI, that doesn't go away.

Lives 24/7 on your VPS
Roams while you sleep
Costs less than a coffee/day

Pricing

One subscription. Everything wired in.

Inference is billed at-cost on top of your plan. No markup, no surprises.

StarterOne agent. Your first home.

$0/mo

1 agent
Private VPS · 1 vCore · 4 GB · 50 GB NVMe
5,000 memories
Pick any open-source model
Inference billed at-cost
Community support

Questions, answered.

Hermes agents are general-purpose autonomous LLM agents. They chat, use tools, run code, browse, and call APIs. You set the system prompt and tools available — alcione provides the runtime.

Ready

Youragentis
0 seconds away.

No credit card. No setup. Create an organization and watch your first Hermes agent come online.

GiveyourteamaHermesagent—with a permanent home.