Operant AI Shipped a Semantic Firewall Plus a Token Meter, Turning Agent Security and Agent Cost Control Into the Same Product
Operant AI launched four products on August 27: Semantic Firewall, which returns an allow/block/redact verdict on every prompt, tool call, command and data movement inline via four guards (Tool Intent, Code Intent, Data Intent, Scope); Token Meter, which applies the same real-time read to spend and enforces budgets mid-session on the specific agent, team, user or model running hot; Browser AI Protection for agentic browsers holding authenticated sessions; and expanded Claude coverage that traces plan steps, tool calls, file operations and sub-agent spawns inside hosted Cowork sessions. The architectural claim worth stealing is that runtime intent inspection produces the cost meter for free, which collapses two SaaS categories (agent observability vendors and FinOps dashboards) into one control plane.
↳ Follow the thread