White Circle, a platform for monitoring and controlling AI models in production, has raised $11m in Seed funding backed by senior figures from OpenAI, Anthropic, DeepMind, Hugging Face, Mistral, Datadog, and Sentry. Founded by Denis Shilov after he demonstrated a single prompt that bypassed safety filters across all major AI models, the platform acts as a single-API control layer that scans AI inputs and outputs in real time. It detects harmful content, catches hallucinations, blocks prompt-injection attacks, flags model drift, and identifies abusive users. The platform supports 150 languages, is SOC 2 and HIPAA compliant, and has served over one billion API requests. Customers include Lovable and two major digital banks. White Circle also released two research pieces: CircleGuardBench, a benchmark for AI moderation models, and KillBench, which found biases and collapsed refusal rates in structured-output production deployments across 15 AI models.
Sort: