SUPERWISE v1.27: Self-Service Starter Tier + Full Guardrails

The Starter edition is free with limited resources. You provide basic details and are ready to go. It includes runtime guardrails, real-time observability, and community support so you can govern AI without a credit card.

SUPERWISE guardrails are designed for using public models safely. They are rule-based policies that combine one or more rules to keep LLM outputs safe and accurate. You define them independently from agents, manage version history, and use a playground to test under different inputs and outputs before publishing. You can block when a rule is violated or let it through and audit it—all with visibility into usage. See docs.superwise.ai/docs/guardrails-1.

MCP (Model Context Protocol) is the standard for exposing agent capabilities. SUPERWISE supports remote MCP with configurable header parameters for authentication so you can connect and govern MCP-based agents in one place.

A playground lets you test how rules apply to different values before going live. Governed AI requires that you verify behavior in advance and revisit published versions. SUPERWISE gives you that control so violations are caught in design, not production.

Yes. SUPERWISE guardrails are designed for using public models safely. SUPERWISE also supports privately hosted and open-source models so customers can build sophisticated AI solutions that handle PII and PHI while keeping data in their control.