Test builds and file reproducible bugs
Connect
SPIRITT
CuaA computer fleet.
One shared mission.
SPIRITT uses Cua to provision pools of computers, distribute desktop work across them, and run parallel jobs with per-run verification. Start with one environment or coordinate an entire fleet.
What you can do with Cua.
Provision fleets
Build reusable pools of task computers from the supported images and runtimes your work needs. Prepare desktop applications and inputs before assigning jobs.
Coordinate parallel work
Distribute GUI workflows across claimed computers with explicit concurrency and ownership. Keep each job tied to its environment, inputs, and output records.
Verify at scale
Collect results across the fleet, run independent success checks, and isolate failures for investigation. Track completed work separately from computers that merely started.
Works with your other apps.
Get started in three steps.
01
Open your fleet workspace
Create a SPIRITT workspace for the desktop jobs, application tests, or agent evaluations you want to run across computers.
02
Connect Cua and choose capacity
Configure Cua Fleet access, choose supported computer images, and set up the pools, applications, and permissions needed by your workload. An existing desktop can also be connected through Driver.
03
Delegate the batch
Hand over the job queue and success criteria. SPIRITT coordinates claims and workers, checks each result, and handles claim release and pool cleanup according to your policy.
Hand over a mission.
Choose a mission. Make it yours.
Put Cua to work.
Put a whole computer fleet to work
Questions
Cua, with SPIRITT.
01What can SPIRITT coordinate with Cua Fleets?+
SPIRITT can build workflows around reusable computer pools, claims, task assignment, parallel workers, and per-run evidence. Cua supplies the computer infrastructure; SPIRITT implements the queue, coordination rules, output checks, and operational interface for your workload.
02Can I still use a single desktop?+
Yes. Cua Driver can operate an existing computer with its required permissions, while Sandbox and Fleets provide separate task environments. A fleet is useful for parallel jobs or isolation between workloads. Background interaction on a personal desktop remains application- and platform-dependent, so dedicated computers are the better fit when your own session must not be interrupted.
03How are jobs distributed across the fleet?+
The workflow can claim computers from configured capacity and assign jobs using stable identifiers, explicit ownership, and a concurrency limit. SPIRITT can build a persistent task queue and operator dashboard, with separate workers executing desktop actions. Retry and duplicate-prevention behavior is implemented for the job rather than assumed.
04Which operating systems can the fleet use?+
Cua describes Fleet and sandbox coverage across Linux, Windows, macOS, and Android. Driver supports macOS, Windows, and Linux desktop operation. Actual availability and behavior depend on the image, runtime, applications, and access configured for your deployment; a test plan should use the environments you have verified.
05What happens to a computer after a job finishes?+
A workflow can retain the artifacts, release its claim, and apply the chosen cleanup or expiry policy. Releasing a claim and deleting Fleet pool capacity are different lifecycle actions. SPIRITT makes those decisions explicit so reusable capacity, retained evidence, and resource cleanup follow your operating policy.
06How do I know the whole batch really completed?+
Each job can have independent checks against its output file or application state, plus a record of inputs, environment, screenshots, and outcome. A batch report distinguishes verified completions from failed or partial runs. Failed jobs can be investigated or rerun from defined starting conditions without treating a started computer as a finished task.