Credits, not executions
One predictable unit for AI spend. Credits map to model usage so finance sees cost before runs complete — not after the bill arrives.
Pricing
Pay for what you run, not what you hope
Seat pricing covers your workspace, governance, and collaboration. AI usage runs on credits — metered before every model call, logged to a tamper-evident ledger, and never a surprise on the invoice.
- $0to start building
- 20%off with annual billing
- 100%governance on every plan
Governance included
Policy gates, provenance chains, and approval workflows ship on every tier. Security is the floor, not an add-on.
Scale on your terms
Upgrade, downgrade, or cancel anytime. Start solo on Free, grow into Pro, and graduate to Enterprise when compliance demands it.
Choose your plan
Plans that grow with your team
Free
Best for solo builders getting started
$0/mo
For solo builders exploring governed AI agents.
- 1 work environment
- Built-in org roles
- 1 seat
Pro
Best for growing teams that need seats + SSO
$100/seat/mo
For growing teams that need SSO and more capacity.
- 3 work environments
- SSO / SAML (built-in roles)
- 25 seats
Max
Best for production teams prioritizing envs + RBAC
$150/seat/mo
Production scale for focused teams — more environments and custom org RBAC. Seat packs available via sales.
- 10 work environments
- Custom org RBAC + SSO
- 10 seats (packs via sales)
Enterprise
Best for regulated orgs and custom deployment
Custom
For regulated organizations with custom governance needs.
- Unlimited work environments
- Custom org RBAC + SCIM
- Unlimited seats
Pro is seat-led for growing teams. Max is environment- and RBAC-led for focused production crews — additional seats via sales. All paid plans include credit-based AI usage. Upgrade, downgrade, or cancel anytime. Enterprise includes custom deployment, residency, and dedicated support.
Compare plans
Feature coverage by tier
| Capability | Free | Pro | Max | Enterprise |
|---|---|---|---|---|
| Work environments | 1 | 3 | 10 | Unlimited |
| Seats | 1 | 25 | 10 | Unlimited |
| SSO / SAML | — | Yes | Yes | Yes |
| Custom org RBAC | — | — | Yes | Yes |
| SCIM provisioning | — | — | — | Yes |
| Credit-based AI usage | Yes | Yes | Yes | Yes |
| Provenance & policy gates | Yes | Yes | Yes | Yes |
| Support | Community | Community | Community | Dedicated |
| Deployment | SaaS | SaaS | SaaS | SaaS / BYOC / on-prem |
Every plan includes
- Governed agent builder
- Provenance audit trail
- Runtime policy enforcement
- Strict knowledge boundaries
- Org roles & permissions
- Community support
How billing works
One credit model, total clarity
Every agent run passes a pre-call credit check. Declined off-policy requests spend zero tokens. Approved runs write idempotent usage events — replay-safe, audit-ready, and transparent in your ledger.
- Build — Design and validate agents on any plan — zero AI required to ship the canvas.
- Check — Runtime verifies credits and policy before the first token is spent.
- Run — Executions meter usage in real time with hash-chained provenance.
- Bill — Credits reconcile to a transparent ledger your finance team can trust.
How credits map to usage
- Typical agent turn — Credits scale with model tier and token length — lighter models and shorter prompts cost less than frontier models.
- RAG-grounded answer — Retrieval plus generation meters as one governed run. Off-policy or ungrounded declines spend zero credits.
- Multi-step workflow — Each model call in a graph is checked before spend; failed policy gates never burn tokens mid-run.
- Finance-ready ledger — Usage events are idempotent and tied to your tenant rate table — finance sees cost before runs complete.
Illustrative of how usage maps to credits. Actual rates depend on model routing and your tenant rate table.