Limina

Limina is the new name for Private AI, the Toronto-based data privacy company founded by Patricia Thaine. Its platform identifies and transforms personal, health, and payment-card information across 50+ entity types and 52+ languages, using transformer-based NLP models. The core design choice is that everything runs inside the customer's own environment, whether on-premises, in a private cloud, or on a major cloud platform, so data never leaves the customer's control and is never shared with the vendor. Deployment is via Docker containers, and the company also offers a cloud API for teams that prefer a managed path.

Beyond detection and redaction, Limina replaces sensitive values with synthetic or context-preserving replacements so de-identified datasets stay useful for analytics and AI training; the September 2026 release 4.5 added Contextual Replacement, a hybrid LLM-plus-deterministic system that keeps replacements realistic (related names stay related, structured values keep valid formats). Named users include Providence Health and Boehringer Ingelheim, which use the platform to unlock restricted data for analytics, partnerships, and AI development. Accuracy and outperform-everyone claims on the site are vendor-reported; the meaningful check is running their evaluation toolkit on your own data. The rebrand from Private AI to Limina is recent, so older documentation and marketplace listings may still carry the old name.

1. Contact Limina at getlimina.ai for the Docker container images or Cloud API access.
2. Deploy the container in your own environment (on-prem, VPC, or cloud).
3. Point your text, document, image, or audio pipelines at the API for detection and redaction.
4. Configure replacement strategies (redact, synthetic, or Contextual Replacement) per entity type.