AgentX
Sign in Start free
Start free
Get started

Start free. No credit card required.

Open a control room in minutes. Your first month includes 1,000 messages; after that Free settles to 200 / month. Upgrade any time. Founder-direct support while you shape the product.

  • No card on file. Create an account and send — billing stays optional until you upgrade.
  • One pooled meter. The whole team draws from the same message pool. Seats never multiply it.
  • Bring your own models. Claude, ChatGPT, OpenAI, OpenRouter — we never bill inference.
Get started → Compare pools

Unlimited on every tier — including Free.

The only number that climbs is your monthly message pool. Everything else stays maxed.

Agents
∞
no cap
Machines
∞
no cap
Sessions
∞
no cap
Concurrency
∞
no cap
Backends
all
included
Models
BYO
your keys
Message pools

Pick the pool size your team needs.

One meter: the messages your team sends. A message is one instruction or reply — nothing else counts, not tokens, not files, not storage. Annual billing is 2 months free on every paid tier.

Free Start here
200 messages / month 1,000 in month one · then the ongoing Free grant
$0/mo $0/yr Billed monthly Billed annually · 2 months free
Start free
Standard
5,000 messages / month
$20/mo $200/yr Billed monthly Billed annually · 2 months free
Choose Standard
Pro Most popular
15,000 messages / month
$40/mo $400/yr Billed monthly Billed annually · 2 months free
Choose Pro
Pro Scale Top tier
100,000 messages / month
$200/mo $2,000/yr Billed monthly Billed annually · 2 months free
Choose Pro Scale
Teams · multi-user

Seats unlock people — not more quota.

Invite teammates, assign roles, and drive one shared fleet. Each seat is a flat $20 / month and draws from the tenant’s single pooled message quota. Seats never multiply the pool, and we never meter per head.

$20/ seat · mo
Pooled quota · roles · one fleet
How the meter works

A message is one instruction or reply — the only thing we meter.

01 · Counts

Instructions and replies.

A message is one instruction or reply passing between you and your agents, or between two agents. Nothing else counts: not tokens, not files, not storage.

02 · Free always

The whole fleet.

Agents, machines, sessions, concurrency, and every backend stay unlimited on every tier — Free included.

03 · Inference

Your models, your bill.

AgentX never bills for AI inference. Keep Claude, ChatGPT, OpenAI, or OpenRouter — we charge for the control room, not the tokens.

Exhaustion New sends pause until the next monthly window. Running work finishes. Nothing is force-stopped.

Sold direct

Need your own private setup — or machines we run for you?

Two enquiry-only offerings, sized and contracted with you. Annual for Enterprise.

Enquiry · annual

Enterprise

Your own private setup, run for you and kept separate from every other customer — with SAML SSO, SCIM, audit retention, a DPA + MSA, and an SLA.

Enquiry

Hosted Runtime

We run the machines for you — no self-hosted daemon to operate. Capacity sized to your fleet.