Agent Hook for Claude Code (public beta, available now), Community Edition (public beta, early September 2026, free), Enterprise Edition (Q4 2026)
Free: $0 forever with 1 agent per engine (up to 3 when friends refer), 1 GB memory. Pro: $12/month or $99/year per engine with 10 agents and 25 GB memory. Unlimited: $39/month or $299/year per engine with unlimited agents and 100 GB memory. Enterprise: from $6,000/year per install on your own server
Key features
Policy-based rule definition for agent actions
Checkpoint enforcement before every action
Tamper-evident signed audit ledger
Vendor-neutral across models and harnesses
Deterministic decision-making without LLMs
Offline enforcement without network connectivity
Budget caps and cost control
Data tokenization and secret placeholder support
One-click installation for Claude Code
Connect ChatGPT, Claude, Claude Code and Cursor with a few clicks
AI tools save decisions, notes, people, customers and files
Search with semantic understanding, not just exact words
Automatic summaries by day, week and month
Link related memories automatically
Prevent duplicate memories with deduplication
Export all data as Markdown and JSON
Separate engines for clients with no data crossing
Fine-grained access control (read or read-and-save)
Full history tracking for every change
What makes it different
Neutral layer enforcing policies across any agent, model, or harness
Self-hosted on your hardware with no telemetry or cloud dependency
Cryptographically verified audit trail you can check offline
Keeps AI tools from forgetting context between sessions
Data isolated per engine with no sharing between customers
Fully portable—export anytime or delete permanently