AQ for startups
Published July 28, 2026 · by the AQ team
AQ lets a startup ship like a team several times its size: run Claude Code, Codex, and other agent CLIs in parallel workspaces on one always-on machine, see everything the agents are doing in one dashboard, and put working previews in front of the founders before anything merges. The Free plan is the full product on a VM you already know how to rent.
The agent bottleneck at a startup is rarely the model. It is everything around it: sessions that die with a closed laptop, work scattered across three engineers' terminals, and a founder who wants to see the feature, not a diff. AQ is the harness that removes those.
Parallel by default
Each piece of work gets its own workspace: an isolated git worktree with its own branch, terminal, editor, and preview. One engineer can drive several agents at once without them stepping on each other, and the whole team can see all of it. When your roadmap is longer than your headcount, this is the multiplier: more work in flight per person, with the isolation handled by the platform instead of by discipline. The mechanics are in git worktrees for coding agents.
Founder feedback without a deploy
Every workspace can run its dev server and expose it as a live preview with a shareable link. A cofounder clicks the link, uses the feature the agent just built, and pins comments directly on the preview. Those comments can be sent to the agent as its next prompt. The feedback loop that used to be "deploy to staging, write it up in Slack" becomes "click, comment, agent fixes it."
Always-on without ops
Agent sessions run server-side, so they keep working when laptops close and the team goes home. On the Free plan that server is a VM in your own cloud account: AQ installs its runner and the agents execute there, with code and CLI logins staying in your custody. When you would rather not think about the VM at all, the Team plan ($100 per user per month promotional, standard $200) includes a dedicated always-on VM that AQ provisions and operates. Both plans are the full product; compare them on pricing.
Costs that scale with people, not tokens
AQ never meters model usage: agents run on the Claude and OpenAI plans your engineers already pay for, with no markup. AQ's price is per seat. For a five-person startup that means the math fits on a napkin, and staying on the Free plan while you validate costs exactly your VM bill.
Frequently asked questions
Do we need a platform engineer to run AQ?
No. On the Free plan you point AQ at a VM in your cloud account and it installs what it needs. On the Team plan there is no VM to manage at all: a dedicated always-on machine is provisioned for you when you subscribe.
Is the Free plan actually the full product?
Yes. Multiplayer workspaces, every supported agent CLI, live previews with comments, and the Linear and GitHub integrations are all in the Free plan. The only difference is who operates the compute: you bring a VM, or the Team plan includes one AQ runs for you.
What happens when we grow past a couple of engineers?
Nothing structural: invite teammates and they get seats in the same workspaces. On the Team plan, each newly invited teammate's seat is free for its first 14 days, so trying a bigger team costs nothing up front.
Can non-engineers use AQ?
Yes, and at a startup they usually do: a founder or designer opens the workspace preview, tries the feature, and pins comments that flow back to the agent. They never need to touch a terminal.