Recon

The fastest, most effective way to rigorously red team your AI apps.

See Recon in action

Request a demo

Scalable Red Teaming for AI

Deploy Your AI Confidently

Recon provides your team the insights to make data-informed decisions about potential risks before you launch AI apps to production. Along with vulnerability assessments, Recon also gives actionable insights for model selection and evaluation of your AI defenses.

Accelerate AI Testing

Recon systematically tests and retests AI apps in a few hours, across multiple threat vectors with an attack library of 450+ known attacks on AI systems, an AI Agent scan, and the ability to bring your own custom library. You can leverage Recon on its own or use it to supercharge your existing red team.

Stay Ahead of Attackers

Recon leverages a community of over 17,000 security researchers, as well as first-party and academic research, to identify and prepare for new threat vectors. With weekly updates to our attack library, threat research is always driving the product, so you know that you are building securely no matter what the latest model or threat is.

Key Features

Extensive Attack Library

Recon’s extensive Attack Library spans six major threat categories and is continuously updated to counter zero-day and emerging attack techniques. Using trained LLMs as detectors, Recon delivers accuracy, not just coverage, to ensure your systems remain protected against a wide range of evolving vulnerabilities with limited false positives.

Protect AI-Site-Product-Recon-Feature-1-042325

Precise, Relevant Red Team Attacks

Recon tests the AI app, not just the model. Recon’s Agent creates relevant attacks that leverage a range of inputs, from the business objective of the application to the base model, deployed guardrails, RAG pipelines, and relevant system prompts.

Augmented Red Teaming

Red teamers can work collaboratively with Recon, using Natural Language to set attack goals—no code necessary. Recon produces in-depth, conversation-level visibility to support risk analysis and remediation.

Protect AI-Site-Product-Recon-Feature-3-042325

Bring-Your-Own Attack Prompts

Enhance your security evaluations with the ability to run custom attack prompt sets. Users can upload bespoke prompts to simulate specific threat scenarios, powered by our proprietary detection algorithms that precisely reflect your environment’s unique needs.

Protect AI-Site-Product-Recon-Feature-4-042325

Reporting and Framework Mapping

Recon easily exports results to CSV and JSON to enable quick and effective collaboration across teams. Vulnerabilities are also mapped to standard frameworks such as OWASP Top 10 for LLMs, DASF, and more, making it fast and easy to meet compliance standards.