Fortinet launches FortiAIGate and FortiGuard Inline AI Security Service for LLM, agent and MCP traffic inspection
Fortinet announced the FortiGuard Inline AI Security Service, a cloud-delivered capability that extends AI-specific inspection and guardrails to existing FortiGate customers. The service is powered by FortiAIGate, which can also be self-hosted on-premises or in customer cloud environments on Kubernetes with NVIDIA GPUs for sovereignty requirements. Capabilities include real-time inspection of prompts, model responses, MCP tool calls and agent actions, shadow AI discovery, DLP for 35 types of PII and confidential data, and spend/rate limits to counter Denial-of-Wallet attacks. The offering incorporates technology from Fortinet's recent acquisition of Virtue AI. This is a product announcement with no incident, vulnerability or active threat disclosed.
Detalhes técnicos
- Multilayered guardrails across prompts, responses, MCP tool calls/responses, API activity and agent actions, covering text, images, audio, video and code.
- DLP can alert, deny or redact sensitive data before exposure.
- Model permissions, budgets, spend and rate limits by organization, workspace, project or user; intelligent routing by content, policy and cost.
- Over 50 predefined compliance policies with detailed audit logs of every prompt, output and tool call.
- Visibility into AI agents and their interactions with enterprise applications, APIs, tools, data and MCP servers.
Linha do tempo
-
Fortinet launches FortiAIGate and FortiGuard Inline AI Security Service for LLM, agent and MCP traffic inspection
ler a análise desta fonte ocultar
Fortinet announced the FortiGuard Inline AI Security Service, a cloud-delivered capability that extends AI-specific inspection and guardrails to existing FortiGate customers. The service is powered by FortiAIGate, which can also be self-hosted on-premises or in customer cloud environments on Kubernetes with NVIDIA GPUs for sovereignty requirements. Capabilities include real-time inspection of prompts, model responses, MCP tool calls and agent actions, shadow AI discovery, DLP for 35 types of PII and confidential data, and spend/rate limits to counter Denial-of-Wallet attacks. The offering incorporates technology from Fortinet's recent acquisition of Virtue AI. This is a product announcement with no incident, vulnerability or active threat disclosed.
Technical Details
- Multilayered guardrails across prompts, responses, MCP tool calls/responses, API activity and agent actions, covering text, images, audio, video and code.
- DLP can alert, deny or redact sensitive data before exposure.
- Model permissions, budgets, spend and rate limits by organization, workspace, project or user; intelligent routing by content, policy and cost.
- Over 50 predefined compliance policies with detailed audit logs of every prompt, output and tool call.
- Visibility into AI agents and their interactions with enterprise applications, APIs, tools, data and MCP servers.