Skyflow, a privacy-focused data protection platform, has introduced a new data protection layer tailored for the Model Context Protocol (MCP), addressing a rising concern for enterprises and SaaS platforms adopting agentic AI. The new platform aims to secure sensitive information as AI agents increasingly interface with real-world tools via MCP.
Originally developed by Anthropic and now backed by OpenAI, AWS, and Google, MCP simplifies how AI agents connect to databases, SaaS platforms and internal systems. However, this standard also opens up fresh risks, particularly around exposing personally identifiable information (PII), health data and financial records.
Skyflow’s solution introduces a polymorphic data protection engine that applies masking, tokenisation and contextual rehydration dynamically, based on usage policies and permissions. This ensures data privacy without disrupting AI performance.
Two deployment models available are the Skyflow MCP Gateway, which acts as a privacy-enforcing proxy, and a Skyflow MCP Server SDK for direct integration into MCP implementations. Key features include audit trails for regulatory compliance, entity-preserving transformations to support AI reasoning and secure memory handling.
“As AI agents start connecting to more real-world data through MCP, companies need privacy infrastructure that can keep up,” Anshu Sharma, CEO of Skyflow, said.
“Skyflow helps developers and SaaS platforms protect sensitive data without slowing down AI workflows, making secure, compliant AI deployment possible at scale.”
The launch extends Skyflow’s existing AI-focused privacy offerings, including its GPT Privacy Vault (2023) and Agentic AI Security and Privacy Layer (2024). The company says the new MCP Data Protection Layer is suited for sectors such as finance, healthcare, travel and retail, where regulatory compliance and data sensitivity are paramount.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.