UnlimitedA mini server of your own, hosted in our Jakarta data centre
Self-learning Agents. Unlimited AI tokens.
Unlimited tokens
Unlimited Agents
Unlimited projects
Agents that learn your business as they work, and never meet a meter. A physical mini server reserved for you in our data centre runs nemotron-3.5-lightning around the clock for a single price: no per-token bills, no monthly fee, nobody else on the hardware.
Self-learning Agents that remember what you teach them
Pay once, use it without limits
Agents at work 24/7, nothing counted
Live within 24 hours
7-day return window
Unlimited tokensNode One · Jakarta · online
one-time purchase, yours outright
Rp25.000.000
$1,399
Node OneUp to 50 Agents at once
CPU
AMD Ryzen 7 7735HS, 8 cores / 16 threads
GPU
AMD Radeon Graphics, 12 cores
RAM
32 GB LPDDR5
Storage
128 GB NVMe SSD
Tokens
Unlimited
Agents
Unlimited
Projects
Unlimited
Model
nemotron-3.5-lightning
Concurrent
1
Monthly fee
None
Your Agents create with the best models for text, image and video
NVIDIA Nemotron
Nano Banana
GPT Image
Kling
Veo
Sora
Flux
Seedream
Claude
Gemini
Midjourney
Why unlimited
Stop paying for every token.
Metered AI punishes the one thing you want: more work done. A mini server turns the bill upside down. One price, and every message, routine and overnight run after it is already paid for.
Pay-per-token AI
The bill grows with every message your Agents send
Shared GPUs, so busy hours slow your team down
Rate limits decide how much gets done today
Your prompts and documents sit on somebody else's cloud
Your mini server
One price. Every token after it is free
Unlimited Agents and unlimited projects, no per-seat pricing
Hardware reserved for you alone, no other workload beside yours
Agents work all night and all weekend at no extra cost
Data stays on your node and is never used for training
Runs nemotron-3.5-lightning one request at a time. Queue as much work as you like; none of it is counted.
What you get
A whole AI team in one console.
Everything your Agents do lives in five tabs: Chat, Project, Agents, Team and Apps. It feels like a messenger because it is one, on desktop, on your phone and on WhatsApp.
Chat
Message your Agents like colleagues
Hand over a task in a chat, on any device. The Agent works through it on your mini server, shows each app action as it happens, and comes back only when it needs your approval.
One thread per Agent, with read receipts and typing like WhatsApp
App actions show up as chips in the bubble, so you see what it did
Anything that moves money or changes access waits for your tap
Grace
Chief of Staff · morning brief
Prep me for the 10:00 with Helix Health.
08:10
On it. Reading the last three threads, the Q3 notes and two open tickets.
08:17
Brief is in your inbox. One thing needs you: they asked for 15% off on renewal.
09:24
Approve reply to Helix Health
Offer 10% on a two-year term, decline 15%.
ApproveEdit
Grace is working…
Message Grace
Running on your mini server · Jakarta data centre
Project
One project per business, client or launch
A project holds its own Agents, team, apps and context: the goal, the website, how you work. Switch projects and every Agent knows which business it is working for. Start from 25 ready-made templates or from scratch.
Templates for agencies, online shops, clinics, SaaS and more
Invite colleagues by email with admin or member roles
The goal and description feed every Agent as context
Projects
Template
NC
Nusantara Coffee
5 Agents · 3 members
Launch team
Gmail, Notion, Slack
In use
HH
Helix Health
3 Agents · 1 member
Support desk
Zendesk, Sheets
SV
Studio Vantage
4 Agents · 2 members
Content
Mailchimp, LinkedIn
CC
Cobalt Capital
2 Agents · 1 member
Finance
Jurnal, Gmail
Agents
Give every job its own Agent
Inbox, invoices, support, outbound, hiring: each gets an Agent with a name, a job and a few lines on how you like things done. Pick a model, connect its apps, set routines and permissions, and read its full history of calls.
Start from a template or a blank Agent in minutes
Per-Agent model, memory, routines and permissions
Every model call and app action logged with its cost
Agents
New Agent
Grace
Chief of Staff
Working
Henry
Invoice Coordinator
Needs approval
Oliver
Inbox Manager
Working
Mia
Support Specialist
Idle
Liam
Recruiter
Paused
Team
Put Agents in a group and let them pass work around
A team is a group chat of up to eight Agents. Mention one, or ask the whole room, and they answer in turn, each seeing what the others said. Your mini server is not metered, so a long group thread costs you nothing extra.
@mention one Agent or address all of them at once
Each reply comes from its own Agent with its own model
Group threads sit in the same chat list as direct ones
Q4 launch
You, Grace, Jack, Ava
Ava
Two competitors shipped pricing pages this week. Summary and screenshots attached.
08:10
Jack
Thanks. October issue updated with a short pricing comparison. Ready for review.
08:17
Grace
Queued for your review at 09:00. Holding the send until approved.
09:24
Approved. Send Friday 08:00.
09:31
Message Q4 launch
Apps
Signed in to the tools you already use
Connect Gmail, Slack, Notion, HubSpot, WhatsApp and more than 1,600 other apps with one sign-in per Agent. Each Agent gets only the apps you give it, and every action it takes shows up in its chat.
OAuth sign-in or an API key, per Agent, never shared
Telegram, Slack, Discord and WhatsApp as messaging gateways
Disconnect an app at any time from the Agent's settings
Apps
Search 1,600+ apps
Gmail
Google Calendar
Slack
Notion
HubSpot
Google Sheets
GitHub
WhatsApp
Zendesk
Linear
Mailchimp
Telegram
5 connected to Grace+1,600 more
How it works
From purchase to first Agent in a day.
No hardware to install and nothing to configure on your side.
01
Buy the mini server
Pay once by card or bank transfer. We reserve a physical node for you in our data centre and send your private access within 24 hours.
02
Create a Agent
Give it a name, a job and a few lines on how you work, or start from a template. It is ready to message in minutes.
03
Hand off the work
Message it like a colleague from the console, your phone or WhatsApp. It comes back with the job done, or a question when it needs you.
Add a Node Cluster when you want models served from hardware reserved for you. Otherwise Agents call any API you bring a key for.
Give each Agent a job
Work you can hand to a Agent.
Start from a template, then tell it how you like things done. Every one runs on your own node.
No plans to compare and no meter to watch. Buy the mini server once, and your Agents run on nemotron-3.5-lightning without limits for as long as you own it.
Unlimited
Mini server
A dedicated mini server in our data centre, yours outright.
one-time purchase
Rp25.000.000
$1,399
CPU
AMD Ryzen 7 7735HS, 8 cores / 16 threads
GPU
AMD Radeon Graphics, 12 cores
RAM
32 GB LPDDR5
Storage
128 GB NVMe SSD
Usage
Unlimited
Agents
Unlimited
Projects
Unlimited
Model
nemotron-3.5-lightning
Concurrent
1
Hosting
Included
Unlimited tokens, no meter ever
Unlimited Agents and projects, no per-seat fee
Nobody else on the hardware
No monthly fee
7-day return window
Live within 24 hours
Add more mini servers to the same private fabric later
Paid in full by card or bank transfer. Prices are in rupiah; the USD figure is for reference.
Customers
Less prompting. More finished work.
“I used to check on the Agents every fifteen minutes. Now the Inbox Manager drafts, the Chief of Staff briefs me at 08:00, and I approve from my phone.”
“Legal would never let a cloud agent into our CRM. A Agent on our own node, in a data centre we can visit, got through review in a week.”
“We showed the Invoice Coordinator the process once. It has run every morning since, and only pings us when a total is off.”
FAQ
Questions we get before the first Agent.
Anything else? An engineer, not a sales bot, answers every message within one business day.
Is it really unlimited?
Yes. The mini server is yours, so nothing on it is metered: no token count, no monthly fee, no overage. It handles one request at a time, so Agents queue their work and get through it around the clock. The limit is the clock, not your bill.
What exactly is a Agent?
A Agent is an AI teammate with a job: inbox, support queue, invoices, outbound, hiring. You message it like a colleague. It has its own desktop, browser and terminal on your node, signs into the tools you already use, and finishes the work, coming back only when it needs your approval.
Where does a Agent run?
On your node: a physical server reserved for you in our own data centre. Nothing is virtualised or shared with another customer. You reach it over a private network, and you can visit the rack.
How many Agents can I run?
As many as your node carries at once: Node Lite is rated for a handful, Node One for a team, Node Pro for a whole company. Add nodes to the same private fabric when you outgrow one; nothing in your Agents changes.
Which models do the Agents use?
Node Lite, One and Pro run Agents, not models. Agents call models over the private fabric from a Node Cluster you own, or from any API you bring a key for. A Node Cluster comes with open-weight chat, code, embedding and vision models preloaded.
Can I talk to a Agent from WhatsApp?
Yes. Every Agent has a chat in the console, and you can connect WhatsApp, Telegram or Slack as a gateway with its own allow-list. The same thread continues on every surface.
What happens to my data?
It stays on your node. Prompts, documents and memory are never used for training, and our staff cannot read them without a support ticket you approve. When you release a node, the drives are cryptographically erased.
How do I pay?
You buy the mini server, paid in full by card or bank transfer. You own the node, we host and manage it in our data centre, and nothing your Agents do is metered. Every purchase has a 7-day return window.