Guardrail Technologies

What it does.

registry ID ndock:tool:guardrail-technologies

Features

• AI Control Panel and Trust Layer: Centralizes configuration of models, prompts, data sources, and agents in one workspace, sitting between users and underlying LLMs so teams can enforce consistent policies across many tools.
• Context-preserving data masking: Replaces sensitive inputs with safe, context-aware aliases instead of blunt redaction, so models still get useful signal while actual identifiers, secrets, and IP stay in the customer’s domain. Authorized roles can later unmask values when permitted.
• Prompt Protect and policy rules: Scans prompts for confidential information, policy violations, or risky language, then blocks, rewrites, or routes them according to customizable rules while preserving enough context for the AI to remain effective.
• Granular role-based access control: Applies fine-grained permissions to who can view, send, or unmask sensitive data, aligning access with job functions and reducing both insider risk and accidental exposure.
• Audit trail and real-time risk intelligence: Logs every prompt, response, and user action, allowing security teams to review incidents, run investigations, and surface alerts when activity deviates from policy or appears fraudulent.
• Model and cloud agnostic integrations: Works alongside Microsoft, Google, OpenAI, Anthropic, Oracle Cloud Infrastructure, and more, so organizations can standardize protection across different vendors and workloads without giving up choice.

Pricing history.

auto-diffed on every sync: the changelog other directories don't keep