Limits
Subbasis limits are designed to make governed execution predictable across plans and deployments.
What this means
Subbasis plan design separates:
- governed execution capacity
- resource limits (users, agents, storage, traces, nodes)
- model usage economics
The Free plan is free forever and usage-limited.
Governed Execution Credit
A governed execution credit measures meaningful execution inside Subbasis: workflow steps, agent tasks, connector calls, approvals, evidence writes, or controlled runtime actions.
Credits are meant to represent governed runtime work, not only token volume.
Common limit dimensions
Depending on plan/deployment, limits can include:
- users
- agents
- workflows
- governed execution credits
- skill profiles
- context/evidence storage
- trace ingestion
- max trace size
- ARIN nodes
- marketplace access scope
Model usage is separate
Model usage is not identical to runtime usage.
- managed model credits can have separate plan/add-on rules
- BYOD keys can be supported under policy constraints
- local model usage depends on deployment profile
This separation helps teams optimize governance and cost independently.
Add-ons
Add-ons may expand execution, storage, observability, local execution capacity, or managed model envelopes depending on current offering.
Limits and deployment notes
- Exact numeric limits vary by plan, release stage, and deployment mode.
- Do not assume every advanced capability is available in every plan by default.
- Validate limits for production operations before rollout.