We host the sandboxed workflow compute and the control plane. There is no worker fleet to provision, patch, scale, or keep alive. Suspend a workflow for a year and it costs storage, not CPU.
Best for the smallest operational footprint
The engine runs in your environment under a commercial license, with enterprise identity, compliance controls, and a team behind it. Dev, staging, and cold-standby clusters are free. Only production counts.
Best when policy or control requires your infrastructure
The execution engine for evaluation and small deployments: sandbox, event store, dispatch, snapshots, replay, all five SDKs, the console, and SQL visibility. It runs on local disk with bearer-token auth, in clusters of up to three nodes. S3 and tiered storage, OIDC and attestation, larger clusters, and multi-region are commercial.
Cloud bills the resources a team already budgets for: CPU, memory, stored data, and network. Workflow shape never appears on the invoice.
Cloud bills active compute and memory, hot and cold storage, and network egress at the provider's published rate for your region. There is no per-workflow, per-step, or per-event charge, and no base fee. A suspended workflow costs storage alone.
Production-scale capability: S3 and tiered cold storage, clusters beyond three nodes, multi-region, OIDC and provenance attestation, and the compliance pack. It also adds a commercial license, named support, and upgrade assistance.
No. Hopskip Cloud hosts the sandboxed workflow compute and control plane. Activities that must reach a private network run through the co-process agent or webhook gateway in your environment.
Cloud is in early access. Intended rates are published now and may move before GA. Early-access customers get a 12-month rate lock at GA. No SLA or compliance certification is claimed before it exists.