Pricing
AQ has two plans. Free: a ready sandbox. Team: your own cloud, your whole team. Both include the real product.
Free
A personal sandbox, created for you. Solo. Nothing to install.
- A machine that's just yours. Once you're in, AQ creates a private dev machine in an isolated network. It sleeps when you step away and wakes with every terminal exactly where you left it.
- Every agent CLI. Claude Code, Codex, Cursor Agent, Kimi, Grok, Antigravity CLI, signed in with your own accounts. Plus the editor, live previews, and Linear and GitHub integration.
- No meters on your models. Your agents run on your own Claude and OpenAI plans; AQ never marks up usage.
Team
/ month
Billed monthly. Join now and your rate is locked for your first 12 months.
- Your whole team. Unlimited seats, shared workspaces, guest links, and public preview links for reviews.
- Your own cloud. Connect VMs in your own account for full custody: code, worktrees, and CLI logins never leave your infrastructure.
- Or we run it. A dedicated, always-on machine in its own isolated network, created with one click once you subscribe. It never sleeps.
- Risk-free team growth. Invite anyone; a seat is only billed once the person accepts and joins.
Neither plan marks up your models: your team's agents run on your own Claude and OpenAI accounts, or on AQ Models with AQ credits priced at provider cost during early access, with $5 included at sign-up.
Which plan should we pick?
Pick Free if you're one person who wants to run coding agents seriously without setting up anything: AQ creates your sandbox and you're working in minutes. Pick Team when you want teammates in the same workspaces, links you can share for review, machines in your own cloud account, or a dedicated machine that never sleeps. You can start Free and upgrade the moment any of those matters.
How is model usage billed?
Two ways, and AQ marks up neither. Most teams sign into the agent CLIs with their own accounts (your Claude subscription, your OpenAI plan): AQ does not sit between you and the model, there is no meter, and the seat price is the whole price. Alternatively, run OpenCode or Pi on AQ Models: that inference consumes AQ credits, priced at provider cost during early access, with $5 in credits included when you sign up. No effort-based pricing on either path. Coming from a credits-based tool? See AQ vs Replit and AQ vs Lovable.
What does "sleeps when idle" mean?
Your free sandbox pauses itself when nothing is running and nobody is at the keyboard, and wakes in about half a minute when you come back: terminals, sessions, and files exactly as you left them. Team machines are always-on and never sleep.
Why is the Team price promotional?
AQ is early and moving fast, and early teams shape the product. The early-access rate is how we say thanks for that: a fraction of the standard price while we build the platform out in the open. Teams that join during the promotion keep the $50 rate for their first 12 months, locked from the day they subscribe.
Larger team?
For bigger deployments (dedicated infrastructure, procurement, security review), email [email protected] and we will put a founder on the call.
Frequently asked questions
Is AQ really free?
Yes. The Free plan is a personal sandbox: AQ creates a private dev machine for you, with the agent CLIs, editor, previews, and integrations included, and no seat charge. It's for one person; teammates, sharing, and your own cloud are what the Team plan adds.
Is there a free trial of the Team plan?
The Free plan is the trial: you evaluate the whole product in your sandbox first, with no time limit. On the Team plan, invitations are free until accepted, so inviting someone who never joins costs nothing.
What counts as a seat on the Team plan?
A person who signs into AQ: engineers running agents, and PMs or designers who join workspaces, comment on previews, and steer work. If they log in, they are a seat.
Does running agents as a team cost more than running them solo?
Usually less per unit of work. One always-on VM hosts the whole team's workspaces instead of an idle machine per engineer, and workspaces are git worktrees that share the repo rather than full per-person environments. Model usage stays on each teammate's existing Claude or OpenAI plan, or on prepaid AQ Models credits (provider cost during early access), and AQ charges per seat with no token markup, so putting a reviewer or a PM in the loop adds no model spend.
Do I need my own Claude or OpenAI subscription?
Only for Claude Code and Codex: you connect those CLIs with your own accounts, which keeps you in control of model spend, and AQ never marks up your usage. Without a subscription, you can run OpenCode or Pi on AQ Models instead: that inference consumes AQ credits, priced at provider cost during early access, with $5 in credits included when you sign up.
Does AQ run in our cloud?
On the Team plan, yes if you want: connect VMs from your own cloud account and code, worktrees, agent sessions, and CLI credentials live there. Or let AQ run a dedicated machine for you. On the Free plan, your sandbox is a private AQ-managed machine in its own isolated network; nothing runs on shared infrastructure on either plan.
What happens when the promotion ends?
Nothing sudden, and everything stays billed monthly. Teams that subscribe during early access pay $50 per user per month for their first 12 months from the day they join. After that, the rate becomes $100 per user per month, still half the $200 standard, and we will tell you well before anything else changes.
Is there special pricing for larger teams?
For larger deployments (dedicated infrastructure, procurement, security review), email [email protected] and we will put a founder on the call.