Durable execution platform used to run stateful agent loops that call models and tools, wait on human decisions for days, and resume exactly where they stopped after failures. It supplies reliability primitives and full execution history rather than AI-specific policy, review consoles or data controls.
hybrid · generally available · Research snapshot 2026-09-06
Demonstrate an agent workflow that pauses for multi-day human approval via a Signal, survives a worker crash, and replays its full execution history.
Capabilities and evidence
Support labels reflect the supplied research. Documentation and vendor claims are not independent product tests. “Not established” means the researcher did not find support; it does not prove a capability is absent.
Documented by provider
Supports long-running stateful agent loops that call LLMs and tools, wait on humans, and resume automatically after a crash, network timeout or multi-day wait.
Limit: Documentation covers execution reliability, not model or data governance.
Approval workflows block until external approval decisions are made, using Signals to capture approval data with metadata, and external sources including human actions can interact with workflows.
Limit: There is no described approval UI, reviewer role model or policy engine; approvals are implemented by developers.
Task queue priority and fairness can prioritise urgent work and keep one tenant from starving others, and worker configuration is used to route and isolate agent workloads.
Limit: This is scheduling fairness, not a security tenant-isolation boundary.