Priced in the open. Scales with your team.
Every product has a transparent entry price — no "contact us for a quote" wall. Enterprise tiers are custom-scoped for multi-region, sovereign, and regulated deployments.
Starling
Kubernetes platform & internal developer platform
Single cluster · up to 10 developers
- Self-service golden paths
- Argo CD GitOps + drift detection
- Shrikeops pre-flight security scans
- Community support
Multi-cluster · multi-cloud
- Everything in Startup
- Multi-cluster control plane
- Istio service mesh + workload identity
- 4-hour support SLA
On-prem · air-gapped · regulated
- Everything in Enterprise
- Air-gapped / on-prem deployment
- Dedicated architecture support
- Custom compliance controls
Reflexion
Self-correcting Actor/Critic agents — FRAKMA ML/LLM Ops
Single / dual cluster · 3 agents
- Actor/Critic Reflexion loop
- Sub-10s streaming RCA
- GitOps-gated execution
- Observability + secrets/TLS included
Unlimited agents · 4-hour SLA
- Everything in Starter
- Unlimited agents
- Cross-incident pattern learning
- 4-hour support SLA
Multi-region · white-glove
- Everything in Growth
- Multi-region deployment
- White-glove onboarding
- Dedicated SRE partnership
Murmuration
AI advisory — strategy, architecture, roadmaps
4-week sprint · 40 hours
- AI architecture assessment
- Cost & risk review
- Prioritized roadmap
- Readout with your leadership
12-week program · 120 hours
- Everything in Discovery
- Hands-on implementation
- Team enablement & training
- Bi-weekly steering cadence
Retainer · 40 hours / month
- Continuous technical advisory
- Architecture review on demand
- Roadmap maintenance
- Priority access to the team
Brain
Intelligent operations — model serving & GPU scaling
Model serving · single GPU pool
- Model serving infrastructure
- Serverless GPU scaling
- MLflow tracking & registry
- Community support
Distributed training · multi-model
- Everything in Build
- Distributed training (KubeRay)
- Multi-model smart routing
- 4-hour support SLA
Multi-region · dedicated capacity
- Everything in Scale
- Reserved GPU capacity
- Multi-region serving
- Dedicated ML platform support
The Startup Pack — 50 seats across the platform at 20% off list.
Multi-region, sovereign, or regulated?
Enterprise deployments are custom-scoped — air-gapped clusters, dedicated capacity, custom compliance controls, and a named architecture team. Typical payback is 2–4 months.
Book an architecture call