RedLens AI

RedLens AI runs automated adversarial attacks against production AI systems and maps findings to controls used in security audits. Its Master Attack Schema v2.0 covers six attack categories and 22 methodologies, and findings are scored by CVSS. Assessments produce PDFs with risk scores, vulnerabilities, remediation steps, and framework mappings; the site describes those mappings as evidence rather than certification. Users can schedule daily or weekly retests to detect drift, with logged email alerts for critical findings. A Python CLI and GitHub Action support pipeline scans with severity build gates and SARIF output, while findings can stream to Splunk, SentinelOne, Devo, or an HMAC-signed webhook. Deployment options include hosted SaaS and Enterprise self-hosted on-premise Docker. In the self-hosted option, customer data is stored in the customer's Postgres, but attack generation and judging still call Anthropic's API using the customer's key. RedLens has a free plan; paid plans include Team at $199/mo, Developer at $799/mo, Professional at $8,000/mo, and Enterprise at $150k/yr+.

RedLens AI combines adversarial scanning, reporting, and scheduled monitoring for production AI. Buyers should note its plan costs, the external Anthropic API calls in its self-hosted setup, and the listed unavailable capabilities.

RedLens AI is aimed at healthcare, law enforcement, and regulated enterprise organizations that deploy AI systems. Its pipeline scans, monitoring, framework mappings, and self-hosted option may suit teams with security audit workflows.

RedLens AI is an automated adversarial-testing and monitoring platform for production models and agents. It suits regulated healthcare, law-enforcement, and enterprise security teams that need repeatable evidence for deployed AI. The strongest case is continuous testing connected to development and security operations; the main drawback is the cost of serious production coverage.

Overview

RedLens attacks production AI and scores the resulting findings with CVSS. Its Master Attack Schema v2.0 organizes six attack categories and 22 methodologies, giving security teams a consistent assessment structure rather than a collection of one-off probes. Reports are delivered as PDFs containing a risk score, vulnerabilities, remediation guidance, and mappings to named security and AI frameworks.

Those mappings include HIPAA, NIST AI RMF, OWASP GenAI Top 10, MITRE ATLAS, ISO/IEC 42001, FBI CJIS Security Policy, and AIUC-1. They are useful as evidence in an audit process, but they are not certifications; compliance teams still need their own attestations and controls.

Key features

Daily or weekly retests are designed to expose production drift. Critical findings can generate logged email alerts, which supports an auditable escalation trail, although Slack and Teams channels are not currently available. A Python CLI and GitHub Action bring scans into CI/CD, with severity-based build gates and SARIF output for GitHub code scanning. That makes RedLens more practical for teams that want security checks to block releases instead of living in a separate dashboard.

The Blast Radius Mapper accepts tool schemas in OpenAPI, Anthropic, or OpenAI formats, covering agent tool surfaces as well as model behavior. Findings can stream to Splunk, SentinelOne, Devo, or an HMAC-signed webhook, providing paths into existing security operations. The product targets AI agents, models, patient chatbots, diagnostic AI, internal copilots, and custom systems.

Deployment options span hosted SaaS, Enterprise self-hosted on-premise Docker, and an MSSP console with per-client isolation. In the self-hosted package, prompts, results, findings, users, and reports remain in the customer's Postgres database. However, attack generation and judging still use Anthropic's API with the customer's key, so on-premise deployment does not remove that external processing dependency. RedLens says its own SOC 2 Type II audit is scheduled to begin in Q2 2027, with certification targeted for Q4 2027; its Railway and Anthropic infrastructure providers are described as SOC 2 Type II certified.

Pricing

The Free plan costs nothing and includes 25 hosted scans per month, a five-scan daily limit, one AI asset under test, and the Claude Haiku 4.5 judge. Work email verification is required. It is suitable for a small proof of concept, but the asset and quota caps make it unsuitable for broad production programs.

Team costs $199 per month and raises capacity to 300 hosted scans monthly, up to 25 scans per day, with no cap on AI assets and a full-strength judge model. It fits teams evaluating multiple assets, while still imposing a hosted-scan ceiling.

Developer costs $799 per month and covers one AI agent or model, one monitored production workflow, monthly automated scans, and single-methodology scheduled scans. It is the lowest paid entry point for ongoing monitoring, but its narrow asset and workflow allowance limits larger deployments.

Professional costs $8,000 per month, supporting up to 25 monitored production workflows, 50,000 attack simulations monthly, weekly automated scans, and unlimited AI assets under test. This is aimed at established security programs rather than occasional assessment.

Enterprise starts at $150,000 per year ($150k/yr+) and removes platform caps on workflows, assets, and simulations. It adds daily automated scans, self-hosted on-premise Docker, and a custom SLA. An Enterprise SLA is available for organizations that require contractual service commitments.

Platforms

RedLens is available through API, web, and self-hosted deployment. The self-hosted option is the meaningful choice for customers needing customer-controlled storage and on-premise Docker, while API and web access suit hosted operations and pipeline integration.

Who it's for

Choose RedLens for regulated organizations deploying patient-facing, law-enforcement, or enterprise AI that need recurring adversarial assessments, framework-oriented reports, and integration with CI/CD and SIEM workflows. It is a poor fit for teams seeking broad employee shadow-AI governance, model-file scanning, content moderation, or DLP: those capabilities are not currently available.

Pros and cons

Pros

• Six attack categories and 22 methodologies provide structured coverage for production assessments.
• CVSS scoring, remediation guidance, PDF exports, and framework mappings turn findings into audit-ready work items.
• Daily or weekly retesting, build gates, SARIF, SIEM streams, and HMAC-signed webhooks connect testing to operational security.
• On-premise Docker and MSSP client isolation support sensitive or multi-client deployments.

Cons

• Production monitoring becomes expensive quickly: Developer is $799 per month, Professional is $8,000 per month, and Enterprise starts at $150,000 per year.
• Self-hosting keeps core data in customer Postgres but still requires Anthropic API calls for attack generation and judging.
• Email is the available critical-alert channel; Slack and Teams integrations are absent.
• There is no current Splunkbase app, PyPI CLI package, broader DLP, shadow-AI governance, model-file scanning, or content moderation.

Alternatives

For a broader directory of options, see AI Red Teaming Tools.

F5 BIG-IP APM is a paid platform with a free trial and coverage across mobile, desktop, API, self-hosted, and web environments; choose it when access-management breadth matters more than RedLens's AI-specific attack and audit workflow.

NVADER offers a freemium web product with a free plan for one app and app scanning, making it a lower-commitment starting point.

Rogue is a freemium self-hosted option with free personal and internal use; choose it when self-hosting and that licensing scope are the priority.

AgentSeal has a free plan with 30 basic probes, an MCP registry, Guard, and Watch, while Pro adds 311 probes, OAuth runtime MCP scanning, and PDF export.

Confident AI provides a free plan with two seats, one project, five weekly test runs, and 1 GB-month of trace spans; it is a better fit for evaluation workloads with modest usage.

Darkhunt AI Security offers a freemium platform with free credits, full platform access, community support, and no card requirement.

Giskard offers a free open-source library with local deployment, basic LLM vulnerability scanning, and a basic RAG evaluation report.

OpenSecureAI Scanner has a $49/month Pro plan with 50,000 API requests, hosted remediation and LLM Gateway completions, and priority email support.

Verdict

RedLens AI is a strong choice for regulated security teams that need continuous adversarial testing, CVSS-based reporting, framework evidence, and integrations that fit release and SIEM processes. Look elsewhere if your budget cannot support production monitoring, if Anthropic API dependence conflicts with your deployment policy, or if you need workplace AI governance and content controls beyond model and agent security.