Pricing
Simple, transparent pricing.
Start free. Scale when you're ready. Deploy on your own NVIDIA hardware at enterprise.
Free
$0/ month
For individual developers and small experiments.
Up to 5,000 session-steps / month
- Persistent sessions (Redis-backed)
- Exact replay engine
- OpenTelemetry tracing
- Hybrid model routing
- OpenAI-compatible proxy
- Community support
Most Popular
Pro
$29–$299/ month
For growing teams running agents in production.
Usage-based — scales with your workload
- Everything in Free
- Unlimited session-steps
- Simulate mode (zero-cost A/B testing)
- Anomaly detection & loop prevention
- GPU utilization & ROI dashboard
- Priority email support
Enterprise
$8k–$60k/ year
Self-hosted on your own NVIDIA hardware. Full data sovereignty.
Custom SLAs + dedicated support
- Everything in Pro
- Self-hosted on your NVIDIA DGX / RTX
- NVIDIA NIM microservices integration
- ICMS / BlueField-4 KV-cache hooks
- Docker + Helm deployment templates
- Dedicated onboarding & SLA support
No SDK changes required at any tier. Works with LangChain, CrewAI, AutoGen, LangGraph, and custom agents.
Pricing is indicative — final tiers confirmed at public launch. Join the waitlist for early access pricing.