Agentic AI Platform Change Log

Integrations

TrueFoundry integrated DeepKeep's AI Firewall into its AI Gateway as a custom guardrail.

Enterprise teams can now enforce advanced security policies and protect against prompt injections or malicious inputs at the gateway level before they reach the agent or model.

Integrations

TrueFoundry added governance and portability support for Claude Sonnet 4.5 and GPT-5 workloads within its AI Gateway.

Enterprises can route, monitor, and secure workloads using these models without needing to rebuild their underlying agent architectures.

Human approval / guardrails

TrueFoundry introduced native human-in-the-loop approval workflows for Model Context Protocol (MCP) tools directly at the AI Gateway boundary. The gateway intercepts designated tool calls, notifies an approver, and suspends the execution state until authorization is manually granted or denied.

Teams can safely connect autonomous agents to high-risk tools and systems by enforcing mandatory human oversight at the infrastructure level. This standardizes the approval process across applications without hardcoding temporary permissions into permanent access policies.

Deployment / data residency

TrueFoundry announced the availability of its AI Gateway on VMware Cloud Foundation (VCF) and vSphere Kubernetes Service. This allows organizations to deploy TrueFoundry's unified model access, policy enforcement, and logging entirely within their own private cloud infrastructure.

Enterprises with strict data sovereignty, regulatory compliance, or air-gapped requirements can now govern their AI operations locally. This eliminates the need to rely on public cloud services for routing and cost attribution.

Observability / auditability

TrueFoundry introduced Ask TFY, a conversational interface built directly into its AI Gateway that acts as an AI agent for platform operations. The interface allows engineering teams to use natural language to query real-time configuration data, analyze traces, troubleshoot bugs, and generate YAML configuration updates.

This feature reduces the time teams spend cross-referencing monitoring dashboards, trace logs, and routing rules across different systems. By allowing operators to diagnose failures and directly apply configuration changes from a single prompt, it lowers the operational overhead of managing complex AI deployments.

Integrations

TrueFoundry has integrated TrojAI into its AI Gateway to enable native real-time traffic enforcement and monitoring. Users can configure the gateway to route traffic through TrojAI as either a proxy layer or a guardrail service to detect prompt injections, data leakage, and unwanted AI content.

This allows security and AI engineering teams to deploy multi-model generative AI applications faster without building custom defense infrastructure. By layering TrojAI threat detection directly onto the TrueFoundry gateway, organizations can block malicious prompts in production traffic natively.

Agent capability

TrueFoundry introduced its Agent Gateway and Agent Harness, providing a dedicated control layer and runtime environment for AI agents in production. The Agent Gateway acts as a data plane for stateful sessions and multi-step executions, while the Agent Harness manages orchestration and context.

Enterprise buyers can now govern agentic traffic, tool calls, and state management across multiple platforms without rebuilding their core applications. This centralizes control over permissions, approvals, and observability, making autonomous agents safer for production deployment.