White Circle raises $11 million for AI control platform that monitors model behavior in company workflows

White Circle, a Paris-based startup founded by Denis Shilov, raised $11 million in seed funding from backers including Romain Huet (OpenAI), Durk Kingma (Anthropic), Guillaume Lample (Mistral), and Thomas Wolf (Hugging Face). The company builds software that sits between users and AI models, checking inputs and outputs in real time against company-specific policies.

The platform acts as a real-time enforcement layer that can flag or block requests for malware, scams, or prohibited content, and can catch model failures such as hallucinations (false information generation), data leaks, unauthorized promises, or destructive actions. White Circle published KillBench in May, a study running more than one million experiments across 15 AI models to test how systems behaved when forced to make decisions about human lives, finding that models made different choices depending on attributes like nationality, religion, or body type.

White Circle says it has processed more than one billion API requests and is already used by Lovable (a vibe-coding startup) and several fintech and legal companies. The company currently has a team of 20 distributed across London, France, Amsterdam, and elsewhere in Europe, and will use the funding to expand its team, accelerate product development, and grow its customer base across the U.S., U.K., and Europe.