We believe every employee should have AI, and every company should keep its secrets.
AI assistants have become part of everyday work. So has the risk of pasting a customer list, a contract or a password into one. Relvant exists so companies don't have to choose between the two.
Our principles
Check before sending, every time
No message reaches an AI model without passing the guard: first messages, follow-ups and regenerated answers alike.
Fail closed
If the security check can't answer, the message isn't sent. Convenience never skips the check.
Our infrastructure, not theirs
The model that reads a sensitive message is one Relvant runs. Only allowed messages go to an outside provider.
Your company sets the rules
We start with sensible defaults. Each company decides what matters to it, and what happens when a rule matches.
Honest about limits
No filter catches everything. We lower the risk with layered checks and measurement, and we never claim leaks are impossible.
A tool people want to use
Streaming answers, history and a choice of models. Security shouldn't make AI worse to use.
Talk to us, or try it yourself.
Write to [email protected], or request access and we will set Relvant up with you.