NeuralTrust

Description

NeuralTrust is a security platform for governing and protecting enterprise AI agents. Its centralized platform lets organizations discover, govern, and protect all their AI agents, with runtime protection, an agent gateway, posture management, and red teaming to enable safer AI adoption at scale.

At its core is TrustGate, an agent gateway that connects company AI agents to the models and tools they need while centralizing security, observability, and governance over every call. TrustGate supports three gateway types in a single control point: an AI Gateway for routing calls across LLM providers such as OpenAI, Anthropic, Azure, and self-hosted models; an MCP Gateway that gives employees scoped tool access with a full call record; and an A2A Gateway for traced agent-to-agent delegation. It forwards end-user identity through every hop, enforces per-agent and per-tool RBAC, maintains cryptographic audit trails, and adds inline prompt inspection, PII redaction, and semantic caching. TrustGate has an open-source core that can be self-hosted in minutes via a single install command.

NeuralTrust separates the control plane from the data plane so customer content stays inside your VPC or on-premises, and it never trains on your data. It supports SaaS, hybrid cloud, and on-premises or air-gapped deployment, is HIPAA and GDPR ready with a FedRAMP and IL4 pathway, and is ISO 27001 certified. It protects both proprietary apps (chatbots, agents, RAG pipelines, APIs) and third-party tools such as Microsoft Copilot, ChatGPT, Claude, and Gemini.

NeuralTrust's Core Features

• TrustGate agent gateway: AI, MCP, and A2A gateways in one control point

• Open-source core self-hostable via a single install command

• Model routing across OpenAI, Anthropic, Azure, and self-hosted models

• End-user identity forwarding with per-agent and per-tool RBAC

• Cryptographic audit trails, inline prompt inspection, and PII redaction

• Runtime protection, posture management, and red teaming

• Control-plane/data-plane separation with no training on customer data

• SaaS, hybrid, and on-premises/air-gapped deployment; HIPAA, GDPR, ISO 27001

How to use NeuralTrust?

1. 1

   Deploy the gateway: Run the open-source TrustGate core locally or self-hosted with a single install command.

2. 2

   Connect agents and models: Point your AI agents at the gateway to route calls across any provider.

3. 3

   Set policies: Enforce per-agent and per-tool RBAC, prompt filters, DLP, and tool permissions centrally.

4. 4

   Monitor and govern: Use full call records, audit trails, and observability across all AI traffic.

5. 5

   Scale up: Upgrade to enterprise for advanced governance, compliance, and multi-team control.

NeuralTrust's Use Cases

• Centralized AI governance
• Secure tool access
• Agent-to-agent workflows
• Regulated deployments
• Red teaming and runtime defense

FAQ from NeuralTrust

• What is NeuralTrust?

  It is an enterprise security platform for governing and protecting AI agents, combining runtime protection, the TrustGate agent gateway, posture management, and red teaming.

• What is an Agent Gateway and why do I need one?

  Without a gateway, each employee secures their own AI traffic or no one does. An agent gateway centralizes security, observability, and governance over every agent call in one control point.

• What protocols does TrustGate support?

  An AI Gateway for routing across any LLM provider, an MCP Gateway for scoped tool access with a full call record, and an A2A Gateway for traced agent-to-agent delegation.

• Is TrustGate open source?

  Yes. TrustGate has an open-source core on GitHub that you can run locally or self-host in minutes, then upgrade to the enterprise version.

• How does NeuralTrust protect data?

  It separates the control plane from the data plane so the data plane runs in your VPC or on-premises and your content never leaves, and it never trains on your data.

• What can I protect with it?

  Both proprietary apps like chatbots, agents, RAG pipelines, and APIs, and third-party tools such as Microsoft Copilot, ChatGPT, Claude, Grok, Perplexity, and Gemini.