Pricing

Price it like Stripe for memory decisions — not like OpenAI for tokens.

You don't pay for decisions. You pay for not making the wrong one.

Save 20% with annual billing

New here?Start with the quickstart →

Free

$0/month

10,000 audited decisions/month (/v1/preflight)

  • ✓ Basic preflight (USE_MEMORY / WARN / ASK_USER / BLOCK)
  • ✓ 4 pillars: Time · Identity · Evidence · Path
  • ✓ Python SDK + MCP server + 6 published framework adapters (3 more in beta)
  • ✓ Community support
  • ✓ 1 API key
  • Plus unlimited /v1/check preview — unmetered on every tier (no audit log)
Start free

No credit card required

MOST POPULAR

Pro

$99/month

250,000 audited decisions/month

Annual: $79/mo (save 20%)

  • ✓ Everything in Free
  • ✓ Full explainability (Omega score, component breakdown)
  • ✓ Repair plan (REFETCH / VERIFY / REBUILD)
  • ✓ Full dashboard + audit log
  • ✓ Webhook + SIEM export (CEF/LEEF)
  • ✓ Email support
  • ✓ Up to 10 agents
  • Hard cap on /v1/preflight audits; /v1/check preview unmetered. Upgrade to Team for 2.5M.
Start Pro

Team

$499/month

2,500,000 audited decisions/month

Annual: $399/mo (save 20%)

  • ✓ Everything in Pro
  • ✓ SSO + RBAC
  • ✓ Unlimited agents
  • ✓ Governance Certificate (W3C VC)
  • ✓ Policy Engine (custom rules)
  • ✓ SLA dashboard
  • ✓ Priority support (Slack)
  • Hard cap on /v1/preflight audits; /v1/check preview unmetered. Upgrade to Enterprise.
Start Team

Enterprise

From $2,500/month

Custom decision limit

  • ✓ Everything in Team
  • ✓ SLA (99.9% uptime)
  • ✓ Dedicated support
  • ✓ On-prem option (+40%)
  • ✓ Custom integrations
  • ✓ BLOCK guarantee: 0 BLOCKs → 20% credit
Contact sales

🚀 Beta pricing — all plans are live today. Save 20% with annual billing. Cancel anytime.

Already have an account? Sign in →

Free tier: 10,000 decisions = enough to protect a full production agent for a week.

All plans include: 6 detection layers · 85-module scoring pipeline · W3C PROV provenance · CVSS-mapped attack surface · A2A protocol support

Two integration paths

Sgraal offers two integration paths: an unmetered preview endpoint (/v1/check) for frictionless experimentation and agent-loop gating — no quota, no audit log — and a metered audit endpoint (/v1/preflight) for production-grade decisions with full audit trail, W3C Verifiable Credentials, and compliance evidence. Both run the same scoring engine; tier limits apply to /v1/preflight audits.

Start protecting your agents today

No credit card required. Demo key: sg_demo_playground

Try in playground Read the docs