Docs Reference
Plans
What each plan unlocks, read from the entitlement table the app enforces.
Every plan badge in these docs comes from this table, which is generated from the same rules the app uses to decide what each plan can do. A self-hosted install without a licence is Community.
Which plan do I need?
Nothing ticked
Community
Every plan includes what Community does.
Limits and features
| Plan | Runs a month | Retention | Seats |
|---|---|---|---|
| Community | 1,000 | 7 days | 1 |
| Starter | 20,000 | 30 days | 5 |
| Growth | 50,000 | 60 days | 10 |
| Scale | 75,000 | 60 days | 20 |
| Pro | 100,000 | 90 days | 25 |
| Enterprise | Unlimited | Unlimited | Unlimited |
| Feature | Community | Starter | Growth | Scale | Pro | Enterprise | |
|---|---|---|---|---|---|---|---|
| See | |||||||
| AlertsRules that notify you by email, Slack, Teams, PagerDuty, Opsgenie or a signed webhook. | From Starter | ||||||
| Fleet overviewEvery agent on one screen. | From Scale | ||||||
| Fleet anomaliesUnusual behaviour across all your recorded runs. | From Pro | ||||||
| Model breakdownError rate, latency and cost per model. | From Scale | ||||||
| Cost breakdownCost by model and by agent, with suggestions to spend less. | From Scale | ||||||
| Peer benchmarksHow your agents compare with anonymised peer workspaces. | From Pro | ||||||
| Team chargebackCost per team, with budget caps. | From Enterprise | ||||||
| Agent mapWhich agents call which, as a graph, with a timeline of each run. | From Scale | ||||||
| Drift detectionSpots a model that starts behaving differently without a change on your side. | From Scale | ||||||
| Trace | |||||||
| Whole-run replayReplay a whole run exactly, clock reads and network fetches included, and score how far a change moves it. | From Enterprise | ||||||
| Multi-agent runsHand-offs between agents stitched into one run, with a graph view. | From Pro | ||||||
| Step replayRun one model call again, with a different model if you like. | Every plan | ||||||
| Stop | |||||||
| Rule libraryReady-made rules from the community, imported in one click. | From Scale | ||||||
| Rule simulationTry a rule against past decisions before you turn it on. | From Scale | ||||||
| Rule impactWhich rules stop which agents most. | From Scale | ||||||
| ApprovalsHold a high-stakes call for a person to approve. | From Starter | ||||||
| Runtime guardRevoke an agent's authority mid-run when a sampled replay shows it misbehaving. | From Enterprise | ||||||
| Fix | |||||||
| Golden suitesRegression tests drawn automatically from your incidents. | From Growth | ||||||
| Model diffHow two model versions behave differently on the same runs. | From Scale | ||||||
| Upgrade gateA CI check that runs your golden suite on a new model before you switch. | From Scale | ||||||
| IncidentsAn incident for each failure, with a suggested root cause. | From Growth | ||||||
| Prove | |||||||
| Audit logA signed audit log for the whole workspace. | Every plan | ||||||
| Compliance reportsReports for SOC 2 and ISO 27001 audits. | From Enterprise | ||||||
| Rule proofsSigned bundles that prove a rule was applied to a given decision. | From Enterprise | ||||||
| Regulatory frameworksYour evidence mapped to named frameworks, such as the EU AI Act and ISO/IEC 42001. | From Enterprise | ||||||
| Hand-off proofsSigned proof of what each agent delegated to the next. | From Pro | ||||||
| Workspace | |||||||
| Team rolesPeople beyond the owner, each with a role. | From Starter | ||||||
| Single sign-onSign in through your identity provider, with directory provisioning. | From Enterprise | ||||||
| Long retentionKeep runs beyond the standard retention. | From Enterprise | ||||||
Prices are on the pricing page.