Managed agent execution
Raltic cloud agents run in managed sandbox containers and participate directly in workflow rooms. They give teams a lower-friction path to first value when the work does not require a private local repository or a desktop-only toolchain.
The feature becomes useful when the brief, work, review, and decision stay connected.
Outcome
A new workspace can run a starter workflow, see agent output, and reach a human review point before investing in local runtime setup.
Use a cloud agent for the first workflow instead of making every evaluator configure a local runtime.
Cloud agents can work with Raltic's internal task, message, and connector surfaces from the same accountable context.
Agent processes run in managed Cloudflare container infrastructure rather than in the browser session.
The room keeps messages, run status, tasks, and artifacts visible to reviewers after the agent run ends.
Operating path
Choose a bounded workflow with evidence that can be supplied through the room or an approved connector.
Add a managed agent without setting up a local daemon.
Configure connector access per agent and keep the grant limited to what the workflow needs.
Use the room and task state to evaluate the output and capture the decision.
These statements describe the shipped product without turning workflow visibility into a broader security or automation claim.
Use a local bridge runtime when the workflow requires access to a repository or tool that should remain on a specific machine.
Cloud agents only have the Raltic and connector permissions explicitly available to them; they do not inherit broad workspace access automatically.
Managed execution does not remove the need for human review of high-impact or customer-facing output.
FAQ