Simple pricing, serious protection

Simple, transparent pricing. Scale as your agents grow. Every plan includes the full guard pipeline — no features locked behind tiers.

Monthly
Annual Save 20%
Starter
Developer
For individual developers and small prototypes.
$20 / month
billed monthly
✓
Up to 10,000 guard calls / month
✓
Up to 5 agents monitored
✓
Cloud API access
✓
Basic dashboard
✓
7-day log retention
—
Custom policies
—
Priority support
Most Popular
Pro
For teams running production agent systems at scale.
$100 / month
billed monthly
✓
Up to 500,000 guard calls / month
✓
Unlimited agents
✓
Cloud API access
✓
Full admin dashboard with alerts
✓
Custom policy engine
✓
Webhook integrations
✓
Email & chat support (24h SLA)
Enterprise
Enterprise
For organizations with strict compliance and scale needs.
Custom
volume-based pricing
✓
Unlimited guard calls
✓
Unlimited agents
✓
Custom detection rules
✓
Air-gapped on-premise + Helm
✓
SSO / SAML / SCIM
✓
SOC 2 & audit log exports
✓
Dedicated support engineer
Detailed comparison
Everything at a glance — see exactly what each plan includes.
Starter Pro Enterprise
Guard pipeline
Prompt injection detection✓✓✓
Sensitive data leakage✓✓✓
Unsafe code generation✓✓✓
System prompt leakage✓✓✓
Custom detection rules—✓✓
Scale
Guard calls / month10K500KUnlimited
Agents monitored5UnlimitedUnlimited
Deployment
Cloud API (SaaS)✓✓✓
On-premise (Docker)——✓
Kubernetes / Helm——✓
Air-gapped deployment——✓
Dashboard & governance
Basic dashboard✓✓✓
Real-time alerts—✓✓
Custom policy engine—✓✓
Webhook integrations—✓✓
Audit log exports——✓
SSO / SAML / SCIM——✓
Support
Community / docs✓✓✓
Email & chat (24h SLA)—✓✓
Dedicated engineer——✓
Frequently asked questions
What counts as a guard call?

One guard call = one input or output passed through the Agent Guard pipeline. A typical agent interaction (user sends message → agent responds) counts as 2 guard calls. Inter-agent messages count separately.

Can I switch between cloud and on-premise?

On-premise deployment is available exclusively on the Enterprise plan. Developer and Pro plans use the cloud API. Enterprise customers can run both simultaneously — the admin dashboard aggregates data from all sources.

Do you store my agent data?

Cloud API mode processes data in real-time and retains only aggregate statistics (threat counts, latency metrics). No prompts, responses, or PII are stored. On-premise mode keeps everything on your infrastructure — zero data leaves your network.

Which agent frameworks do you support?

Agent Guard is framework-agnostic. It works with LangGraph, CrewAI, AutoGen, custom Python/Node agents, or any system that can make HTTP calls. Integration takes 3 lines of code for the cloud API.

What happens if I exceed my plan limits?

We never block your agents. On Developer, excess calls are queued with slightly higher latency. On Pro, overage is billed at $0.0004 per call. Enterprise plans have no limits.

Is there a trial for Pro?

The Pro plan includes a 14-day free trial with full feature access, no credit card required. The Developer plan is available immediately at $20/month.