WorkGroove FAQs

WorkGroove brings product and engineering together to plan, dispatch, and review AI work across teammates and connected computers. Individuals can organize their own work; team plans add shared coordination and governance. It is for developers, engineering leads, founders, and teammates who contribute briefs, plans, and feedback.

Those tools do the AI work. WorkGroove connects the request, plan, progress, decisions, recorded costs, and result so you can follow the whole job. Keep your supported tools and subscriptions. Team plans add assignment to eligible teammates, shared visibility within permissions, budgets, and review where your policies require it.

Developers keep their supported AI tools, subscriptions, repositories, and GitHub workflow. Connect a work computer and start the jobs you want to coordinate through WorkGroove. It keeps those jobs and their results together, with team collaboration subject to permissions. Standalone terminal sessions are not automatically imported into WorkGroove.

Yes. Once a work computer and supported provider are set up, teammates can describe an outcome, add context, and start eligible work from the web console. Team plans support dispatch to eligible teammates. Access, model availability, and review follow your permissions and configured policies; execution needs an online work computer.

Review the available plans and access options, then complete initial setup on a computer. Connect a supported AI provider, authorize the repository or files needed, and describe a small first task with a clear outcome. Keep the work computer online while it runs, then review the result in the console.

Yes. Single builder plans support your own work, recorded costs, and Work Records. Move to a team plan when you need multiple builder seats, assignment, shared Context, shared review policies, or team budgets. Shared access follows membership and permissions; choosing a team plan does not make every project visible to everyone.

Changing plans does not delete your saved work. The active plan controls new capacity and how much history remains available. The Plan page shows current usage before you make a change.

The Free plan is $0 and includes one builder seat, one work computer, 50 usage units per month, five saved AI Assistants, and 90 days of history. Access follows the current availability process. AI provider access and charges are separate; the Pricing page explains capacity and how usage units are counted.

Every plan includes AI Assistants, Workflows, automations, and a Work Record for each run. Capabilities do not have a saved-item limit. Plans change the number of builder seats, work computers, monthly usage, saved AI Assistants, and days of history.

WorkGroove offers a Free plan and paid plans with different capacity and team controls. Your AI subscriptions, API usage, and other provider charges remain separate from WorkGroove billing. Choose capacity for the work you need and review current prices, usage allowances, and billing terms on the Pricing page.

Bring the approved cloud, desktop, or local AI access you already use. Those accounts and subscriptions remain separate from WorkGroove billing.

Personal AI spend reporting is available on every plan. Team plans add shared budgets and organization reporting for authorized roles. Reports distinguish provider-reported costs, estimates, and unpriced usage. Budgets check eligible work before it starts; owners and organization admins bypass managed limits, and budgets do not cap provider invoices.

Budget Off keeps attribution and reporting. Warn reports overages without denying work. An owner or admin prepares and confirms Hard mode, which checks estimates and available capacity before covered work starts. Bypassed usage is reported separately.

AI tools run on your connected work computer. WorkGroove stores conversation history and Work Records in its backend, and processes shared attachments and previews. Published work goes to its configured destination. Selected AI providers and tools may receive task content, so local execution does not mean all data stays on your device.

Your work computer runs the configured provider against the authorized files and repository. WorkGroove’s hosted services keep the request, messages, available decisions, output references, recorded costs, and evidence status so authorized users can follow and review the work.

Submitted attachments and supported previews can be stored and displayed by WorkGroove. Full unpublished artifact files remain on the originating computer; preview and publication state can be hosted. Publishing can send content to configured repository, file, or other publication destinations, where the destination’s access rules also apply.

Instructions, selected files, attachments, repository context, conversation content, and related metadata may be sent to the selected AI provider or connected tools. Their retention and training practices depend on their terms, your account, and settings. See the Privacy Policy for the complete data handling explanation.

Supported tools include Codex, Claude, Muse Code, Cursor, Zen, OpenCode, Gemini (via Antigravity CLI), Grok Build, Nano Banana (Images via Antigravity), Copilot, GLM, LM Studio (Local models). Available models depend on your connected provider, account access, configuration, and work computer. WorkGroove does not include every model subscription or override provider limits.

Claude uses Claude Code, and Copilot uses its supported CLI. Zen model access depends on the connected OpenCode setup. Jev (Beta) is optional model selection and Knowledge curation through TypeSafe, with separate opt-in for each feature; it is not a generative coding provider.

Yes. Execution requires an online work computer with the selected provider ready and authorized. A phone or tablet browser can control work but does not replace that computer. If it disconnects, new work cannot run there until it reconnects; interrupted work may need recovery or another attempt, depending on the provider and session.

Open the web console in a phone or tablet browser to check for waiting questions, respond, and review results within your permissions. Complete initial setup on a computer and keep an execution computer online. Work that needs your answer waits for it. Browser push alerts are not available.

No. WorkGroove keeps the request, available decisions, recorded costs, and results for work started through WorkGroove. Standalone terminal sessions are not automatically imported. Start the work you want to coordinate in WorkGroove so its progress and available evidence can stay together in the resulting Work Record.

You have the produced files or repository changes and a Work Record that connects the request, available plan, decisions, recorded costs, and result. Available checks and missing evidence are shown explicitly. A finished run does not itself prove correctness or release: review the output and recorded publication or apply status.

Signed Work Records export as WGR V1 files for verification against an explicit trust store. Verification covers recorded claims and stated limits; it does not fill evidence gaps or guarantee the work is correct.

Yes. Team plans support shared review policies and named reviewers for governed work. Approval is required where configured policy calls for it, within the reviewer’s permissions. People can review, request revisions, or hold work. Not every run waits for approval, and an approval alone does not prove an external release completed.

Configure the relevant review and publication rules for the workflow and destination. The Work Record keeps captured decisions and available apply or publication evidence. Repository policies and required reviews still apply; check the destination’s result before treating work as shipped.

Yes. Every plan supports Workflows, automations, and ordinary repeat where available. Save work as a Task on single builder plans or a Work Request on team plans. Scheduled execution still needs eligible capacity, permissions, and an online work computer with the selected provider ready.

Exact Replay is limited to eligible workflows on team plans whose saved contract and preflight still verify. Replay verification does not promise byte-identical model output. Ordinary repeat and Exact Replay are different capabilities; the Work Record shows the available path.

Access follows organization membership, roles, and team or repository permissions. Team plans provide shared work and review for authorized people; they do not give every teammate access to every project. Administrators manage membership, and published work may have a different audience under its destination’s access rules.

Authorized administrators, reviewers, and external collaborators can see information within their permissions. Shared work can remain accessible to authorized members after someone leaves. History availability also depends on the active plan and retention terms.

Connect the provider subscription, API access, or compatible local model you are authorized to use. Credential handling depends on the integration: local provider sign-in and configured service credentials have different storage paths. WorkGroove uses protected credential storage and permission checks, and scopes provider and tool access to authorized work.

Review the provider, requested permissions, and account settings before connecting it. Provider terms and billing remain separate. Do not put passwords or API keys in task prompts, shared attachments, or support messages. Contact support for help identifying the setup used by a specific integration.

Use the setup guides in the product docs after signing in, or contact support@workgroove.ai through the Support form. Describe the problem, your computer and provider, and the relevant session if available. Share only information needed to investigate, and leave passwords, API keys, and unrelated private files out of your message.