FOR TEAMS

Small software for
your team’s tasks.

Operations describes a task to the coding AI it already uses. The AI writes the code; runvibe_ provides isolated environments to run the solution under engineering policies. Your team validates the results and shares what can be reused.

Explore small software Explore controls and governance

Small software for everyday tasks.

Small scripts and applications can meet specific operational needs: query data, generate a spreadsheet, check payments received, or update records through an authorized action.

Monthly metrics

Query monthly metrics using a function authorized by engineering.

Tracking spreadsheets

Turn sales files into a tracking spreadsheet for your department.

Authorized actions

Prepare to mark a payment as received or update selected records through a function with defined inputs and permissions.

From verification to results

Finance describes a verification routine to its AI. The solution uses integrations prepared by engineering and, after the required review, can record a payment through an authorized function. The department checks the result.

Engineering prepares the foundations so each department can address specific needs without a developer dedicated to every task.

From a need to a function the team reuses.

Start with a function prepared by engineering or use an authorized environment to create new code. Each path has its own permissions; review applies to the uses defined by policy.

Two ways to collaborate

Use Tools prepared by the team

Call a function through your AI or application and supply the required inputs. Execution uses authorized dependencies without giving the caller direct access to the Image or Secrets.

Build on authorized Images

Use tools and dependencies prepared by engineering. Your AI can create the code for a task; submission and execution follow the applicable policies and permissions.

Explore the approval workflow ↓

Human review for new code, when required

Packages that require approval are blocked from execution until an authorized person reviews the exact revision.

  1. Submit for review. The AI prepares code and submits the files in a Package.
  2. Receive the decision. An authorized person reviews the exact revision and approves or rejects it. Rejection includes a reason.
  3. Make corrections when needed. Your AI can read the reason, adjust the code, and submit a new revision, which requires approval again.
  4. Run with authorization. Only the approved revision becomes eligible; permissions, configuration, and policies are still checked.

The responsible team validates the business result. Approved code can run directly or be published as a Tool when there is a useful function to reuse and share.

Choose seats for your team.

The Teams plan includes three Development seats for US$60 per month, with a shared balance of US$72. Add Operations seats to create and use authorized Tools, or Development seats to also work with Workspace, Images, Executions, and the API.

Compare seats and explore the Teams plan ↗

Use your company identity and control execution access.

Corporate login with SSO

Use your team's corporate identity to access runvibe_ through SSO.

Network access

Choose which destinations each Execution can access. For containers, image downloads and running code have separate network controls.

Environment access

Access Workspaces through the authenticated shell. To test services, forward ports to your computer without exposing them directly to the internet.

Credentials with Secrets

Store credentials in Secrets and reference them in the execution configuration. The process receives values through environment variables, so you do not need to write them into code or the manifest.

Keep executed code separate from other environments.

Each Workspace and Execution runs in a virtual machine with a kernel separate from the host and other machines. This isolation reduces the exposure of the infrastructure and other environments to executed code.

Containers within the same Execution share its virtual machine and kernel.

Virtualization and orchestration

The runvibe_ virtualization layer is based on a Firecracker fork. A proprietary orchestrator manages the virtual machines used by Workspaces and Executions.

Review what ran on the platform.

Use records of operations performed on runvibe_ to investigate failures and review work done with AI or through the shell.

Activities, records, and how to use them in a review
ActivityWhat you can reviewHow to use it
Shell workCommands sent and session outputReview operations performed
Code executionsStatus, output and error logs, recorded metricsInvestigate failures and monitor execution
Generated resultsStructured data and collected filesCheck processing deliverables

Container logs are available after collection at the end of an execution. Results and files follow the platform's limits and retention periods.

Connect your team's tools and company applications.

Use MCP-compatible AI clients or work directly through the shell. Integrate Tools and Executions with your applications and services through the SDK or API.

Codex

Claude Code

Cursor

GitHub Copilot

Kiro

Gemini

Workspace

Develop and test in an environment you can pause and resume. Prepare images to reuse tools and dependencies in compatible new environments.

Explore Workspace ↗

Tools

Define inputs and link a manifest to reuse calls through MCP, SDK, or API. Each call starts an Execution.

Explore Tools ↗

Executions

Run code on demand and access results in your applications, with resources and network access configured for each execution.

Explore Executions ↗

Need to run on your company’s infrastructure?

Explore Enterprise to deploy runvibe_ in your private cloud or on-premises, with installation assistance and platform operation support.

Discover solutions for companies ↗

Frequently asked questions.

Does collaboration also work in the runvibe_ cloud?

Yes. The Teams plan provides collaboration and a shared balance in the runvibe_ cloud. See Pricing for included seats and plan terms.

Who approves AI-generated code?

When policy requires review, an authorized person approves or rejects the exact Package revision. Rejection includes a reason; a correction creates a new revision and requires approval again.

Does approval guarantee the result is correct?

No. Approval records an authorized person’s acceptance of that content. The responsible team still needs to validate the business result.

Can pending or rejected Packages run?

In workflows that require review, pending or rejected Packages stay blocked. This also applies to code called through a Tool.

Do we need a private deployment to work as a team?

The Teams plan supports collaboration in the runvibe_ cloud. To run the platform in your company’s infrastructure, explore Enterprise on the For companies page.

How do I get started with my team?

Choose the Teams plan and fill in the form to tell us how your team plans to use runvibe_. See Pricing for included seats and shared balance options.

Do you also develop products?

Discover runvibe_ for devs ↗

Bring your team to build on runvibe_.

Request access to evaluate collaboration, environments, and controls against your organization’s needs.