Early access: your sandbox is free, with $5 of AQ Composer credits every month. Your own subscriptions stay unmetered. Start free

aq.dev / docs/what-is-aq

What is AQ?

AQ is the multiplayer coding harness where engineering teams run AI coding agents like Claude Code and Codex together: shared live terminals, a code editor, and app previews, in your own cloud. It lives at aq.dev, and the name is short for AgentQueue. The agents are the real CLIs (Claude Code, Codex, Cursor Agent, Devin CLI, Kimi, Grok, Antigravity CLI, or a plain shell), running in persistent sessions on an always-on machine your team controls, streamed live to the browser: close your laptop and the agent keeps working, and any teammate with access can open the same workspace and watch or steer the same session. If someone mentioned AQ in a thread and your search engine sent you everywhere except here, this page clears it up.

The name: AQ, aq.dev, and AgentQueue

AQ is short for AgentQueue, the product's original name. The product, the company accounts, and the website all go by AQ today, and the site is aq.dev. Two-letter names are genuinely hard to search for, so the reliable ways in are typing aq.dev into the address bar, or searching for "aq.dev" or "AQ multiplayer coding" instead of the bare letters.

Not the AQ you were looking for?

The letters collide with a few unrelated things. AQ at aq.dev is not Amazon Q Developer, not an air quality index, and not the adversity quotient. AQ is an independent product built by BetterLeap, Inc. and is not affiliated with Amazon or AWS. The Amazon confusion is also aging out: as of September 2026, AWS has announced end of support for Amazon Q Developer (new signups closed May 15, 2026, IDE plugins and paid subscriptions end April 30, 2027, with AWS pointing customers to its Kiro product instead), while AQ at aq.dev is an active, unrelated product.

What AQ actually does

Each task your team hands to an agent gets a workspace backed by its own isolated git worktree on a branch like ai/1234-fix-auth, so ten agents on ten tasks never step on each other's files. Inside the workspace, the real agent CLIs run in persistent tmux sessions on the team's VM: close your laptop and the agent keeps working, open the workspace from your phone and the same session is there.

The multiplayer part is what the name promises. A teammate with access opens the same workspace and sees the same live terminal, the same editor, the same running preview. Typing into someone else's terminal is delegated: the teammate requests control, the owner approves in one click, and the owner keeps their own input throughout. That makes steering an agent (and reviewing the session) a team activity instead of a solo one. Reviewers pin comments directly on elements of the running app preview, and any comment can be handed to the agent as a follow-up prompt.

Work flows in from Linear or GitHub, and flows out as pull requests that agents open with each user's own GitHub access, tracked per workspace. Model spend stays on the subscriptions your team already pays for: each user connects Claude Code, Codex, and the other agent CLIs with their own accounts, and AQ never marks up usage on your own subscriptions.

Where it runs and what it costs

Agent execution happens on a dedicated machine, never a shared multi-tenant tier. On the Free plan AQ creates a personal sandbox for one person: a private machine in an isolated network, nothing to install, no time limit. On the Team plan, $50 per user per month in early access (standard price $200, billed monthly, with your rate locked for your first 12 months), you connect VMs from your own cloud account or use a dedicated always-on VM that AQ manages in its own isolated network. Inviting a teammate costs nothing until the person accepts and joins. Details are on the pricing page; security and isolation covers how the boundaries work.

Where to go next

The docs overview walks through the moving parts. Multiplayer coding agents explains the category AQ sits in, and harness of harnesses defines the shape of the product: one place that runs every agent CLI rather than another agent of its own. Or skip the reading: aq.dev, Start free.

Frequently asked questions

Which coding agents run on AQ?

Claude Code, Codex, Cursor Agent, Devin CLI, Kimi, Grok, Antigravity CLI, and plain shells. Each runs as the real CLI (not a reimplementation) in a persistent tmux session on your team's machine, signed in with each user's own account, and AQ never marks up usage on your own subscriptions.

Does AQ run in my own cloud?

Yes. On the Team plan you connect VMs from your own cloud account, or use a dedicated always-on VM that AQ manages in its own isolated network. There is no shared multi-tenant execution tier: agent sessions, code, and CLI credentials live on your team's own machine.

How do teammates join a live agent session?

They open the same workspace in the browser and see the same live terminal, editor, and app preview. Typing into someone else's terminal is delegated: a teammate requests control and the session owner approves in one click, keeping their own input. Workspace visibility is owner-managed: team-visible, or private and shared with specific people.

Is AQ free?

There is a free plan with no time limit: a personal sandbox for one person, on a private machine AQ creates in an isolated network, with nothing to install. The Team plan is $50 per user per month in early access (standard price $200, billed monthly), and inviting a teammate costs nothing until the person accepts and joins.

What does AQ stand for?

AgentQueue. The product goes by AQ, and the website is aq.dev.

Is AQ the same as Amazon Q Developer?

No. AQ (aq.dev) is an independent product built by BetterLeap, Inc. and is not affiliated with Amazon or AWS. Amazon Q Developer was Amazon's AI assistant for AWS development; as of September 2026 AWS has announced its end of support (April 30, 2027) and directs customers to Kiro.