Pebblo is an AI security and governance platform designed to enable safe use of enterprise data in AI applications and agents. It provides a suite of components that enforce access control, data classification, compliance, and security across AI workflows — including RAG (Retrieval Augmented Generation) pipelines, AI chatbots, and agentic systems. Core components include: - **Safe Connectors**: Permissions-aware, classifier-enabled data pipelines connecting AI apps to enterprise data sources. - **Safe Retriever**: Ensures only role-appropriate, compliant, and safe data is fetched from vector databases. - **Safe MCP**: Secures unified data access for MCP (Model Context Protocol) agents with identity and policy enforcement, protecting against hidden command injection attacks. - **Safe Prompt**: Removes sensitive data from prompts and prevents prompt injection attacks. - **Safe Inference (Safe Infer)**: Acts as an AI gateway routing requests to policy-approved, authorized AI models or adapters, preventing leakage of sensitive data, code, or IP. - **Data Governance and Security**: Provides monitoring, policy enforcement, and security controls for AI apps, retrieved data, and user behavior. Pebblo is available as both an open-source SDK and a commercial enterprise platform. It targets AI builders, governance teams, and security teams operating in regulated industries. The platform is recognized in the 2025 Gartner Market Guide for AI Trust, Risk and Security Management.