Runtime security layer that stops risky AI actions before they execute
High-assurance infrastructure for AI agents
Stage
Not given
Not given
Founded
Not given
Not given
Based in
Not given
US
Pricing model
Freemium
Open source
Pricing
Free plan: Local Guard runtime with real-time approval prompts and local audit log. Solo $4.99/mo: adds 2 cloud-connected devices and 30-day cloud retention. Pro $15/mo: adds 5 cloud-connected devices, 180-day retention, and policy versioning. Team $30/seat/mo: up to 25 cloud-connected devices, 365-
Not given
Key features
Surface-specific pre-action interception
Policy routing with allow, observe, ask, and block outcomes
Local review and evidence workflows
MCP, skill, plugin, and package inspection
Layered local and team policy
Open source runtime under Apache-2.0
Offline local protection without cloud connectivity
Continuous GitHub leaked-secret monitoring
Decision audit trail and export
Support for multiple AI coding agents
Harpe: chooses what an AI agent reads under a token budget with consistent results
Kibisis: keeps agent memory on your hardware with AES-256-GCM encryption
Argus: append-only record that detects edits with cryptographic verification
No network code by default for local and disconnected use
MIT License with published code, benchmarks and raw results
What makes it different
Open-source core runtime runs locally without cloud requirement
Protects against secrets, destructive commands, prompt injection, and supply chain attacks
No source code or secrets uploaded; decisions stored as redacted summaries
Measured empirical validation across 6,111 questions outperforming full-context models
Zero plaintext on disk with encrypted vaults matching plaintext accuracy
78.3% fewer tokens sent to the model with a 1,200-token budget per turn
Preserves conversational chronology and relational qualifiers under token budgets
Perseus Computing