Enforce Policies

Vera applies custom company policies to the inputs and outputs of models, preventing unwanted content from being sent to Generative AI models.

Vera uses a blend of deterministic, probabilistic, and agent-based techniques to cover a vast risk surface. Vera’s multifaceted approach catches more PII, passwords, topics, and sensitive content by leveraging state of the art moderation that goes beyond a blocklist.

Vera can be used in two ways:

  1. Our flexible API makes it easy to embed filtering and customization into your AI-powered product

  2. The managed chat window gives a ChatGPT-like experience while granting safe and secure access to your approved OSS, custom, and commercially available models

Both experiences provide a consistent, managed experience no matter which model you decide to use, including commercially-available, open-source, and custom models you’ve built yourself.