← Latest products

Shieldstral

Define safety at runtime for text and images. Shieldstral is a 3B open-weight multimodal guardrail from Mistral.

Visit mistral.aiexternal site

Reactions (1)

Similar products

by description similarity

  1. Retroguard

    retroguard.ai

    Verifiably secure AI guardrails

    Hacker News6 pts

  2. Voxtral

    mistral.ai

    Frontier open source speech understanding models

    Product Hunt194 votes

  3. Warestack

    warestack.com

    Agentic guardrails for safe releases

    Product Hunt428 votes

  4. Voxtral TTS by Mistral AI

    mistral.ai

    Multilingual TTS model with realistic and expressive speech

    Product Hunt156 votes

  5. AI Coding Agent Guardrails enforced at runtime

    sigmashake.com

    Generative AI agent guardrails mapped to MITRE ATLAS AML.M0020: deterministic ALLOW, DENY, ASK, FORCE at the tool-call boundary. Honest-agent. Not a sandbox.

    Hacker News5 pts