Skip to content

VBR Labs

Our product engineering group. We take problems we meet repeatedly in enterprise engineering work, starting with AI and MCP security, and build products to solve them.

An AI agent sends tool calls to the VBR MCP Security Gateway, which applies identity, policy, threat detection, data security and audit controls before forwarding them to MCP servers that reach enterprise APIs, databases, SaaS and internal tools.

Current research areas

MCP and agent security
How to authenticate, authorize and audit AI agents that call tools through Model Context Protocol servers.
LLM application threats
Prompt injection, insecure output handling and data leakage in production LLM applications.
AI governance at scale
Discovering and governing AI usage across large organizations with many teams and providers.
Platform intelligence
Joining cost, security and reliability signals so platform teams know what to fix first.

From research to product

Every Labs product goes through the same stages, and we label where each one is.

  1. Stage 1
    Research

    Study a problem we keep meeting in enterprise engineering work and the attacks or failures behind it.

  2. Stage 2
    Prototype

    Build a working prototype and test it against realistic scenarios.

  3. Stage 3
    Design partners

    Run it with a small number of enterprises in their own environments.

  4. Stage 4
    Product

    Harden, document and support it as a product.

Work with VBR Labs

If your teams are adopting AI agents or MCP, we'd like to build with you. Design partners get early access and influence the roadmap.