ProductHermes agent · managed VM
PricingFlat monthly · unlimited use
StatusService online
Cohort3 of 8 slots filled

Your AI assistant. Not by the token.

A Hermes agent on your own VM, with the AI included and unlimited. No credit balance. No upstream rate limit. No model surprises. The same stack we run for live businesses today.

Apply for a slot → See the stack
Founding cohort 3 / 8 filled
37.5%
01 · The difference

The 1-Click product is fine for tinkering. This isn't that.

Hostinger's OpenClaw is built for solo users who want to play with an AI assistant. Alagad is built for a business that needs the agent to handle real customer volume — and to keep handling it without the bill becoming a problem.

The question
Hostinger 1-Click OpenClaw
Alagad
Who controls the agent?
A managed dashboard. You configure what they let you. Custom tools and integrations follow their roadmap.
You do. Self-managed VM. Root access. Your own Hermes config. Plug it into Shopify, Sheets, internal webhooks — anything that speaks HTTP.
What happens when usage grows?
AI credits deplete. You top up. Heavy days cost more. Volume becomes a line item.
Flat monthly rate, AI included. Hammer the agent as hard as your business needs — the bill doesn't move.
Which model is answering?
Whichever upstream provider is cheapest this quarter. Routes change. Quality drifts.
Gemma-4-31B on dedicated infrastructure. Same model, same response quality, every conversation.
Who answers when something breaks?
A support ticket queue.
A founding-cohort operator. Same person who built the reference deployment.
02 · What ships

A VM, an agent, and the AI to run it.

Four discrete pieces. Each one yours. You configure, we operate.

Your VM

A · ubuntu 24.04

A real Ubuntu virtual machine with your own root password and SSH key. Self-managed. Snapshot, restore, customize.

CPU2 vCPU RAM8 GB Disk100 GB NVMe NetworkPublic IPv4

Hermes agent

B · pre-installed

The agent framework from Nous Research. Built-in messaging gateway, learning loop, OpenAI-compatible model client. You configure the personality.

ChannelsWhatsApp · IG · FB · Telegram ToolsHTTP · webhooks · MCP MemorySurrealDB · mem0 LicenseOpen source

AI inference

C · unlimited

An API endpoint scoped to your VM. Use it as much as your business needs — the rate is flat, the AI is included.

ModelGemma-4-31B Context64K tokens Tool callsnative Meteringnone

Control panel

D · self-service

A control panel at app.alagad.net for snapshots, console access, resource graphs, and IP management. No ticket required.

ConsolenoVNC · in-browser Backupsscheduled · on-demand MetricsCPU · RAM · disk · net Auth2FA enforced
03 · The AI is included

No credits to top up. No meter to watch.

The reason a flat-rate plan works is because we run the inference. No upstream API to mark up. No third-party rate limit to inherit. Your subscription pays for the AI — you don't pay again per message.

Inference · scoped to your VM Live · serving
Model
Gemma-4-31B
Quantization
AWQ Marlin
Context
64,000 tokens
Tool calling
native · auto-parse
Streaming
SSE · OpenAI-compatible
Pricing model
flat · unlimited
Routing
LiteLLM · least-busy
Failover
warm secondary
∞
Inference calls per month
Use your agent like an employee, not a meter. Capacity is sized so every founding client can work peak hours simultaneously.
0
Per-token charges
Your subscription buys the assistant. There is no credit balance to deplete and nothing to top up.
8
Maximum founding clients
Capped to keep the service comfortable. The next cohort opens with the next batch of capacity.
04 · Reference deployment

A working example. Handling DMs today.

Mira Le Fleur is a Manila-based florist serving Makati and BGC. Their assistant — built on the same stack we ship to founding clients — handles inbound Instagram and Facebook messages around the clock.

BusinessMira Le Fleur
VerticalFlorist · DTC
ChannelsIG DM · FB Messenger
StackHermes · Gemma-4-31B
StorageSurrealDB · mem0
BridgeBeeper
Status● Production

“The same architecture that handles a flower shop's late-night order inquiries is what we hand to every founding client. Same agent, same response quality, same operator. The only thing that changes is whose business the agent learns.

Alagad · Operator note
05 · The handoff

You bring the business. We keep the lights on.

[YOU]

You handle

  • The personality and prompt your agent uses
  • Connecting your messaging accounts and channels
  • Workflows and integrations specific to your shop
  • Knowing your customers — the agent learns from you
[US]

We handle

  • Provisioning the VM and pre-installing Hermes
  • Keeping the AI service warm and the model loaded
  • Updates, security patches, scheduled backups
  • A direct line to the operator who set up your stack
06 · The terms

One tier. One price. Locked.

We're not running a tier ladder. The founding cohort gets the same VM, the same AI access, the same operator attention — at a rate that holds for twelve months.

Founding cohort · 5 slots remaining SKU · ALG-STD-V1

Alagad Standard

  • 2 vCPU · 8 GB RAM · 100 GB NVMe — your own VM
  • Hermes agent framework, pre-installed and configured
  • AI included · Gemma-4-31B · unlimited use
  • Self-service dashboard at app.alagad.net
  • Onboarding session with the operator
  • 30-day money-back guarantee
Founding rate locked for 12 months. After cohort closes, standard pricing is ₱28,000/month.
Setup and onboarding included for founding clients.
₱22,000 per month · founding rate
07 · Questions

Things people actually ask.

Q.01
What does "AI included, unlimited" really mean?
+
There is no token meter on your account. Your VM has an API key for the inference service, scoped to fair use. We size capacity so every founding client can work peak business hours simultaneously without contention. We monitor utilization and will tell you (and add capacity) long before it ever becomes a problem.
Q.02
Can I use a different model?
+
Gemma-4-31B is the canonical model for this product. It's what runs in the reference deployment, and standardizing on one model keeps the operational story simple and the response quality consistent. Custom-model needs are evaluated case by case after onboarding.
Q.03
What kind of business is this for?
+
SMEs whose customers reach them through messaging — flower shops, salons, clinics, bakeries, boutique e-commerce. The math works best when your agent generates real conversation volume; if you only need an assistant for occasional internal tasks, a per-token cloud product will cost less.
Q.04
What does the onboarding actually look like?
+
A 60–90 minute working session with the operator. We provision the VM, walk through the Hermes config, connect the first messaging channel, and write the initial prompt together. Most founding clients are answering live customer messages within 48 hours of signing.
Q.05
Can I leave?
+
Any time, no penalty. The first 30 days are money-back, no questions asked. After that, monthly billing with no commitment. We'll export your agent's memory and configuration on request.
Q.06
Why eight clients?
+
The service has the capacity to serve more — but eight gives every founding client room to grow without contention, and gives the operator time to actually know your business. The next cohort opens when the next batch of capacity does.

Tell us about your shop.

We respond within 48 hours. If you're a fit, we'll reach out directly to schedule the onboarding session.

Applications close when eight founding slots are filled. We'd rather move slowly with the right businesses than fast with the wrong ones.

Response window ≤ 48h
Onboarding session ~ 90 min
Live within ~ 48h of signing
Money-back window 30 days
Founding · application ● Open