Pricing for the Context Firewall lifecycle.
Start with attack-surface recon, move into Context Firewall Monitor mode, enforce runtime policy for live agents, and add human-led teardowns when you need deeper proof.
Solo builders kicking the tires
- Automated agent attack-surface scan
- Core families: prompt injection, indirect injection, jailbreaks
- 1 scan / week
- Shareable findings report
- Community support
Early teams that need firewall visibility before enforcement
- Everything in Free, plus:
- All 14 attack families
- On-demand + weekly offensive teardowns
- Exploit replay mapped to context-to-action paths
- Context Firewall in Monitor mode
- Generated source-to-sink policy pack
- Email support
Teams ready to enforce the Context Firewall in production
- Everything in Pro, plus:
- Context Firewall in Enforce mode
- Block · redact · gate · quarantine at runtime
- Continuous validation on every change (CI/CD)
- Evidence pack + control matrix export
- Slack, Jira & CI integrations
- SSO + priority SLA
Regulated, multi-agent & MCP-heavy estates
- Everything in Team, plus:
- Human-led offensive teardowns
- Self-host / VPC firewall + tenant isolation
- SAML/SCIM, RBAC & audit logs
- SOC 2 + OWASP-LLM / MITRE ATLAS evidence
- Continuous-validation retainer + SLA
- Dedicated security engineer
Prices in USD. Secure Stripe checkout is being configured for Pro and Team; today, “Get started” reserves your plan and our team activates it within a business day. Enterprise is scoped to your agent count, toolchain, and workflow risk.
One-time, human-reviewed teardown of a single agent, converted into Context Firewall policy. Credits toward an annual plan.
Add protected agents beyond your tier’s included count.
Extra mappings and evidence — SOC 2, ISO 27001, HIPAA — on Enterprise.
Three units, no surprises.
You’re metered on what actually drives risk and cost — agents, teardowns, and the runtime decisions the firewall makes — with hard caps and spend limits so a busy month never means a runaway bill.
One AI workflow or app you test or protect. Most teams start with 1–3 and add agents as they ship.
One automated offensive teardown pass against an agent. Free runs weekly; paid runs on-demand and on every change.
One runtime decision the firewall makes on a request. Pro includes 100k/mo in Monitor mode; Team 2M/mo in Enforce.
Past your included allowance, usage bills at a flat per-block rate — about $40–$50 per 100k inspections or per 1k probes.
Set a hard monthly budget. We alert at 80% and 100%, and never blow past it silently.
Heavy accounts move to usage-based or Enterprise — we never throttle a live security control to protect a flat price.
// inspections = runtime firewall decisions · probes = offensive teardown passes · annual billing saves ~20%
Every control, mapped to a plan.
The firewall capabilities and attack coverage that ship with each tier.
Start by where you are.
Different buyers need a different first step. Pick the one that matches how close you are to deploying a real agent.
Questions, answered.
Start free, or book a Context Firewall Review.
One agent, one workflow. We’ll map the context boundary, replay exploit paths, and show where the firewall should enforce.