SpecWeave Studio

SpecWeave Studio is the workspace for the whole SpecWeave approach. You give a project a goal; a coordinator splits it into tasks and hands them to workers. The coordinator and every worker can use a different coding agent and model: Claude Code, Codex, Grok Build, Cursor, OpenCode, GitHub Copilot and others. They run on machines you own, next to your code, skills and accounts.

Underneath, the open-source SpecWeave CLI keeps the spec, the task evidence and the handoff in git. That record is what lets any agent continue work that another one started.

Checkout recoveryProject · 3 machines

CoordinatorClaude Code · Opus 5.5

Customers lose their cart when they leave checkout. Fix it, and never restore a paid order.

Planned three tasks from increment 0042. T-01 goes to Codex on the Studio Mac, T-02 to Grok Build on the Mac mini, T-03 stays here. I will check each result against its acceptance criteria.

MemoryRelease moved to Friday. All workers see it.

Workersown provider and model

T-01 Save the checkout draftCodex · GPT-6.1-SolStudio Mac · worktreeRunning

T-02 Restore it on returnGrok Build · grok-4.7Mac mini · worktreeNeeds your approval

T-03 Never restore a paid orderClaude Code · Sonnet 5.5This Mac · worktreeVerified · npm test

Plan · SpecWeave increment 0042

1 verified · 1 running · 1 waiting on you

Illustration. Task names, models and machines are examples; your providers and machines decide what is available.

Availability

Studio is in a private pilot and has no public download yet. Sessions, machines, approvals and usage limits are in the pilot build. Projects (coordinator, workers, memory, plans and routines) are being tested for the next pilot build. Ask about a pilot.

What a project holds

Part	What it does
Coordinator	One long conversation per project. It plans, assigns tasks to workers, reviews what they report and asks you when a decision is yours.
Workers	Native sessions, each with its own provider, model, machine and worktree. You can open any worker and steer it directly.
Memory and settings	Per-project memory and instructions. A decision you give the coordinator reaches every new worker.
Plan (optional)	Spec, acceptance criteria, tasks, assignments and evidence. It can be a native Studio plan or read an existing SpecWeave increment. Reported done and Verified are separate: a task is verified only when its saved check runs and passes.
Folders and repositories	A main folder on a machine, extra folders you approve, and several repository URLs. Studio clones only when you ask it to, using that machine's own Git sign-in.
Connections	Local MCP servers per project. Studio discovers their tools and you grant each tool explicitly.
Routines	Scheduled work that runs inside the project.
Personal projects	A project with its own working folder and no Git, for research, notes and everyday tasks.

Pausing and archiving a project are reversible. Deleting one removes its memory, plans and routines but keeps your source folders and remote branches.

How it differs from cloud project agents

Topic	Cloud project agents	SpecWeave Studio
Where work runs	A cloud container per thread, with a bridge to your computer in some products.	On machines you own: this computer, plus any you pair over your network, SSH or Tailscale.
Models	One vendor’s models.	Claude Code, Codex, Grok Build, Cursor, OpenCode, Copilot and more. The coordinator and each worker choose their own.
Your skills and folders	Upload skills to the account or a repository. Neighbouring local folders are out of reach.	The agent runs on your machine with its provider’s local configuration, so local skills, MCP servers and folders you approve are within reach.
When a limit hits	Wait for the reset or change plans.	Switch the thread to another provider, which carries the recent conversation over; resume on the same account when the limit frees up; or hand off to another tool with SpecWeave.
The record	Inside the vendor’s project.	Specs, task evidence and handoffs in git, readable by any tool.
Laptop closed	Keeps running in the cloud.	Runs while the executing machine is awake, so a desktop or Mac mini can carry work while your laptop sleeps.

Cloud project agents include Claude Code Projects and OpenAI Dots. Compared from their public documentation in October 2026; these products change quickly.

The trade-off is simple. Cloud project agents keep running when every one of your machines is off, and they need no setup. Studio needs the machine that does the work to be awake, and in return it uses your local skills and folders, your existing subscriptions from several vendors, and a record that is not locked inside one product.

Machines and networking

Studio runs as a small service on each machine, at localhost:8319, and you use it from a browser. A desktop app for Mac and Windows is in development.

• One machine needs nothing else.
• Other machines on your home or office network can be paired directly once their network access is on.
• A machine you already reach over SSH can be connected through an SSH tunnel that the desktop app opens; it can also start Studio there.
• Away from that network, or from a phone, Tailscale Serve gives private HTTPS to each machine. Studio never uses Tailscale Funnel, which would make it public.

Tailscale is optional. One machine can keep several routes and use the first that answers.

Being reachable is not permission. Every machine and every browser is paired with Studio separately, and native sign-ins for Claude Code, Codex and the others stay on the machine that owns them. See the architecture and its boundaries.

Usage limits and handoff

Studio reads the limits of the accounts on each machine. When you start new work, a capacity policy (Reset first, Balanced or Spend first) picks a fresh eligible subscription. When a conversation hits a limit, Resume when available continues it on the same account once the provider allows it again.

You can also switch a thread to another provider yourself. Studio passes a budgeted part of the conversation to the new agent, which keeps working in the same checkout. Studio does not yet switch on its own when a provider reaches a usage limit. Moving work outside Studio is a SpecWeave handoff: the spec, ledger and work in progress are committed, and the next agent, in Studio or anywhere else, runs specweave pickup. Between turns, automatic checkpoints keep a local recovery point.

Optional pieces

Piece	Do you need it?	What it is for
Tailscale	Only for private HTTPS beyond your local network or from a phone	Private HTTPS transport. Studio pairing is still required.
CLIProxyAPI	No	If you already run a CLIProxyAPI hub, Studio can show its pooled accounts' limits and redeem Codex reset credits. Native sessions never pass through it. Keep the hub on loopback: its management key reaches every pooled account. Check each provider's terms before pooling subscriptions.
AnyModel	No	An adapter for API and local models that a native agent does not support.
Switch	Optional, part of the pilot install	Reads account sign-in and limit observations on each machine for the capacity policy.

Questions

Will my local skills work? Yes, on the machine where the agent runs. Studio starts the provider's own agent there with its configuration, so the skills, MCP servers and settings you set up for it apply. A skill on another machine has to be installed on that machine too.

Is my code sent anywhere new? No. Folders, sessions and project data stay on the machine. Hosted relay, telemetry and self-update are off by default. Model calls go to the providers you choose, as they would from your terminal.

Is SpecWeave Studio the same as Skill Studio? No. npx vskill studio opens Skill Studio, which develops and evaluates skills.