AQ vs Tutti VM: a shared room, or a shared machine

Published September 9, 2026 · by the AQ team

Choose AQ if you want your team's coding agents running on an always-on machine the whole team shares: real CLI sessions on a VM in your own cloud, watched and steered from any browser, alive after every laptop closes. Choose Tutti VM if you want agents to keep running locally on each person's own machine, with a shared cloud Room that syncs everyone's working state so two people's agents can build the same thing at the same time. Both are multiplayer and both are bring-your-own-subscription; they take opposite positions on the most consequential question in the category: where does the agent actually run?

Tutti VM (tutti.sh, styled Tutti · VM by its maker) opened Early Access on August 27, 2026 and is moving fast, so every Tutti claim below is as of September 2026, verified against the Tutti open-source repository on GitHub (tutti-os/tutti) and Tutti's own launch announcement.

AQ vs Tutti VM at a glance

What is Tutti VM?

Tutti VM is the Early Access product of Tutti, a Hong Kong team that also maintains an open-source project of the same name (Apache 2.0, tutti-os/tutti on GitHub). Tutti calls it a "multi-user, multi-agent, real-time collaboration space" and pitched the launch as "the Google Docs moment for cross-agent collaboration." The open-source Tutti app gives one person a shared workspace across the agents on a single machine; Tutti VM extends that across people and devices through a managed cloud Room.

The architecture is its most interesting claim. Your Claude Code or Codex keeps running as a local tool on your own computer, inside a managed local VM. Multi-layer virtualization maps the agent's runtime into a managed Linux environment, intercepting its process calls and file operations and deciding, per instruction, whether each one runs on the physical machine, in the VM, or in the cloud. The result Tutti advertises: login sessions, subscriptions, credentials, API keys, SSH keys, and corporate SSO never leave the machine, while the working state (what is being discussed, built, and produced) lives in the Room, where every collaborator sees it in real time.

Inside a Room, two people's agents can edit the same file at the same time, with conflicts avoided and resolved at the filesystem layer in the background rather than as a merge to untangle afterward. Typing @ references any collaborator's session, file, or output, so one agent can continue another's work without re-briefing, and a collaborator can authorize their agent to be "borrowed" by someone else. A site running on someone's localhost opens directly in the Room with no deploy, and files are shared by link. The app also ships its own group chat, docs, slides, image and video generation, design, task management, and an agent board. Early Access is free and bring-your-own-subscription on macOS and Windows; creating a Room takes an invite code (released in daily batches), while downloading the app or joining an existing Room does not.

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.

On AQ the agents are the real CLIs your engineers already use (Claude Code, Codex, Cursor Agent, Kimi, Grok, or plain shells), running in persistent tmux sessions on the team's VM and streamed live to the browser. Each engineer signs in with their own Claude or OpenAI account, and AQ never marks up model usage. Every workspace gets one isolated git worktree (branch ai/{id}-{slug}, automatic dependency install, one-click rebase onto main), so parallel sessions never collide. Sessions survive a closed laptop and resume from any device, including a phone. Teammates open the same workspace and watch the same live session; typing into someone else's terminal is delegated by its owner approving a control request in one click. Each workspace runs a live dev-server preview with shareable links that work without an account for viewing, and pinned comments on the preview can be sent to the agent as prompts. Label a Linear issue ai-task and a workspace appears; humans start agents, never auto-launched. There is no shared multi-tenant execution tier.

Where Tutti VM stands out

• Local stays local, radically. Login sessions, API keys, SSH keys, and corporate SSO remain on your machine and the cloud never holds them: the strongest form of "my agent never leaves my laptop" that still allows collaboration.
• Same-file simultaneity. Two agents editing one file at the same time, with conflicts handled at the filesystem layer, is a genuinely novel claim; if you want Google-Docs-style co-editing rather than parallel branches, nothing else shipped is shaped like this.
• Cross-agent context. The Big @ reference model (any collaborator's session, file, or output, available to any agent) and borrowed agents attack the re-briefing tax directly.
• An app ecosystem in the room. Docs, slides, image and video generation, and task management ship inside the product, so non-code artifacts live where the agents work.
• No new spend during Early Access. The product is free and rides the subscriptions you already pay for.

Where AQ wins

• The machine outlives the laptop. Tutti VM's compute is each owner's own computer; the Room syncs state, not execution. On AQ the agent runs on an always-on VM, so a session started at 6pm is still working at midnight and resumable from a phone at 7am. See always-on coding agents and keeping Claude Code running after closing your laptop.
• One shared machine, not N synced ones. Everyone on AQ works against the same VM: same dependencies, same environment, same running dev server. There is no per-person environment drift to reconcile, which is the classic failure mode of sharing a dev environment across a team.
• Isolation by construction. AQ gives every workspace its own git worktree and branch, so parallel agents cannot collide in the first place; the history stays ordinary git. Filesystem-level conflict resolution is impressive, but parallel session collisions are a problem we prefer to make impossible rather than manage.
• The terminal is the shared surface. Teammates watch the actual live session, keystroke for keystroke, and the owner can hand over input with one click. Visibility is owner-managed: team-visible, or private and shared with specific people.
• Browser everywhere. AQ needs no desktop install and works on macOS, Windows, Linux, and mobile; Tutti VM is a macOS and Windows app as of September 2026.
• Team plumbing. Per-user GitHub auth, PRs tracked per workspace, Linear intake with two-way status sync, and preview links outside reviewers can comment on after a quick email sign-in.

The deeper difference: syncing machines vs sharing one

Tutti VM and AQ agree on more than most rivals in this space: coding agents should be a team activity in real time, not a queue of finished diffs, and engineers should keep their own subscriptions. That consensus is still rare. A July 2026 LeadDev analysis of 25,264 agent-generated pull requests across 2,361 popular GitHub repositories found that in 79 percent of agentic PRs the same developer both reviewed and modified the agent's contribution, and only about one in eight workflows involved multiple humans. Both products exist to break that pattern; they are two designs for multiplayer coding agents, and both push the industry toward the word multiplayer agents.

The split is architectural. Tutti VM keeps compute personal and makes state collective: your machine runs your agent, and virtualization projects its work into a shared Room. AQ makes the compute itself collective: the team's VM runs the agents, and browsers are just windows onto it. The Tutti design maximizes local custody and lets each person's exact local setup participate. The AQ design maximizes durability and uniformity: sessions that outlive any laptop, one environment everyone trusts, and work that belongs to the team's infrastructure rather than to whichever machine happened to run it. Which one you want depends on whether "the work lives on my machine" reads to you as a security feature or as a single point of failure.

Pricing compared

Tutti VM, as of September 2026: free during Early Access, bring-your-own-subscription (currently Claude Code and Codex), with invite codes required to create a Room and released in daily batches. Post-Early-Access pricing is not published.

AQ has two plans. Free is a personal sandbox for one person: AQ creates a private machine in an isolated network, nothing to install, no time limit. Team is $50 per user per month early access (standard $200, billed monthly), covering VMs you connect from your own cloud or a dedicated always-on AQ-managed VM in its own isolated network, with your rate locked for your first 12 months. Inviting a teammate costs nothing until the person accepts and joins. On both platforms, model usage rides your own accounts: AQ never marks up model usage.

Bottom line

Tutti VM is the closest thing AQ has to a true philosophical rival, and a serious piece of engineering: local agents, local credentials, and a Room where two people's agents co-edit in real time is a design nobody else has shipped. If your priority is that agents and identity never leave each person's machine, take the Early Access invite. AQ's bet is that the machine itself should be the shared, durable thing: agents on an always-on VM in your own cloud, sessions any teammate can open, watch, and steer from a browser, worktree isolation instead of conflict resolution, and work that keeps running when every laptop in the Room is asleep. If that is the failure mode you care about, that is AQ's side of the line.