Pricing
Simple per-seat pricing. Pooled model spend, no markup.
Every plan includes access to every approved model. You pay for people, not for tokens.
Team
For pilots and small orgs
- Up to 50 users
- All frontier models included
- 3 knowledge connectors
- Basic audit log (90 days)
- SSO (Google, Microsoft)
- Community support
Business
For companies rolling out AI org-wide
- Unlimited users
- All 7 knowledge connectors
- Human-in-the-loop approvals
- Immutable audit log (7 years)
- SCIM + custom SSO
- Emergency stop
- Priority support with 4h SLA
Enterprise
For regulated and sovereign deployments
- Everything in Business
- Customer-managed keys (BYOK)
- Data residency (US, EU, UK, APAC)
- Self-hosted / air-gapped option
- EU AI Act Annex IV auto-doc
- Named security engineer
- 24/7 support with 1h SLA
Questions
What buyers usually ask us.
How does model spend work?
Pooled across your organization. We route to the cheapest capable model per prompt and pass through provider costs at no markup on Business and Enterprise.
Do you train on our data?
No. Ever. Tenant-isolated storage, no fine-tuning on your content, and BAAs where applicable with every underlying provider.
What connectors are included?
Google Workspace, Microsoft 365, Slack, GitHub, Confluence, Notion and Salesforce on Business and Enterprise. Team includes 3 of your choice.
Can we self-host?
Yes, on Enterprise. Thyreus runs on your own Kubernetes with air-gapped model providers for defense, intelligence and sovereign workloads.