AI Security
Pebblo
Identity-aware data controls for AI apps and agents, enforced at retrieval and inference time.
Pebblo Overview
What it does
Pebblo is an AI data governance and security platform that lets enterprise AI applications, agents, and coding assistants use company data without exposing information users are not authorized to see. Its distinctive mechanism is the TwinGuard architecture: a Data Ingestion Gateway classifies data by confidentiality, compliance, and semantic context, while a User Access Gateway evaluates every query against user identity, document intent, and application context before results reach the model.
How it works
Safe Connectors ingest data from enterprise sources while extracting each source's fine-grained permissions, and Safe Retriever filters retrieved context at query time so only role-appropriate documents reach the language model, working with existing vector stores. Safe Prompt and Safe Infer inspect prompts and completions inline, redacting secrets and stripping injection payloads, while the Safe MCP gateway vets, sandboxes, and applies runtime permissions to Model Context Protocol servers and tools. Policies are defined in a central policy plane, and audit trails of AI interactions feed real-time alerts. An MIT-licensed open source core integrates with generative AI frameworks including LangChain and LlamaIndex.
Credentials and traction
Included in the 2025 Gartner Market Guide for AI Trust, Risk and Security Management for the AI Governance and Runtime Inspection and Enforcement use case, and named in Gartner's 2026 Information Governance landscape. Pebblo's MIT-licensed open source core is distributed on GitHub and integrates with mainstream generative AI frameworks. The platform targets enterprises deploying retrieval-augmented generation, homegrown agents, and AI coding assistants over proprietary data.
Key Capabilities
mapped to solution categoriesEnforces document-level access at retrieval time so a user receives only context they are authorized to see, filtering before the vector search, after retrieval, or both.
Intercepts prompts and completions to prevent sensitive data (PII, credentials, internal IP), from being transmitted to external LLM services or returned in model responses.
Detects and blocks adversarial inputs designed to override system prompts, extract training data, or redirect model behavior. Detection approaches include pattern matching, input semantic analysis, and secondary model classification.
Enforces IAM-style policies on LLM API access, controlling which users and applications can invoke which models and data sources, with audit logging.
Evaluates model outputs against content policy, data classification rules, and format expectations before delivery to end users, blocking responses containing sensitive data or policy violations.
Secures AI coding assistants and their Model Context Protocol connections against unsafe actions, data exposure and supply-chain risks.
Discovers, governs and allowlists the Model Context Protocol servers and tools that AI agents are permitted to invoke.
Records prompts, completions, and metadata for all AI interactions with tamper-resistant storage, supporting compliance, forensics, and policy investigation.
Automatically discovers AI models, LLM API connections, ML pipelines, and AI-enabled SaaS applications in use across the organization, including those deployed without IT authorization.
Detects sensitive or regulated data in AI training, fine-tuning, or third-party LLM flows without appropriate controls, such as unencrypted PII in inputs or PHI sent to external APIs.
Maps data lineage and provenance across AI training and inference pipelines, tracing how PII, PHI, and IP move into models and external services.
Monitors AI-agent behavior at runtime to detect anomalous or malicious actions and policy violations.
Discovers and enforces least-privilege access for non-human and AI-agent identities across systems and data.
Integrations
compatible toolsImplementation & support
Info last updated on July 4, 2026
Vendors
Is this your product?
Claim your profile to connect with the teams looking for your solutions.