RedLens AI
redlens.ai
- Free plan
- Web
- API
- Self-hosted
What RedLens AI is
RedLens AI runs automated adversarial attacks against production AI systems and connects findings to security audit controls. It is aimed at organizations deploying AI in healthcare, law enforcement, and regulated enterprise settings. Its Master Attack Schema v2.0 covers six attack categories and 22 methodologies, and scores findings with CVSS. Each assessment produces a PDF containing a risk score, vulnerabilities, remediation steps, and mappings to frameworks including HIPAA, NIST AI RMF, OWASP GenAI Top 10, MITRE ATLAS, ISO/IEC 42001, FBI CJIS Security Policy, and AIUC-1. The maker describes these mappings as evidence, not certification. Teams can schedule daily or weekly retests to check for drift, with logged email alerts for critical findings. The Blast Radius Mapper accepts tool schemas in OpenAPI, Anthropic, or OpenAI formats. A Python CLI and GitHub Action support pipeline scans with severity build gates and SARIF output for GitHub code scanning. Findings can stream to several SIEM services or an HMAC-signed webhook. Deployment options include hosted SaaS and Enterprise self-hosted Docker. Self-hosting stores prompts and results in the customer’s Postgres, while attack generation and judging still use Anthropic’s API.
Who it suits
It suits healthcare, law enforcement, and regulated enterprise organizations assessing production AI systems. Teams needing pipeline scans, scheduled retests, or self-hosted deployment may find its listed options relevant.
Good to know, and what to watch
- PDF reports include remediation steps and framework mappings.
- Daily or weekly retests can detect drift.
- GitHub Action supports severity gates and SARIF output.
- Findings can stream to listed SIEM tools.
- Enterprise offers self-hosted Docker deployment.
- Attack generation and judging call Anthropic’s API.
- RedLens SOC 2 Type II audit is scheduled to begin in Q2 2027.
- Slack and Teams alert channels are not currently available.
- Professional is listed at $8,000/mo.
Plans
Every plan and priceEnterprise
Free
Team
Developer
Getting started
- Visit https://redlens.ai/ to review the hosted and self-hosted options.
- Choose a plan: Free, Team, Developer, Professional, or Enterprise.
- Use the hosted SaaS option, or select Enterprise for self-hosted on-premise Docker.
- Connect AI assets or tool schemas to assess; the Blast Radius Mapper accepts OpenAPI, Anthropic, and OpenAI formats.
- For pipeline scans, use the Python CLI or GitHub Action.
- Set scheduled retests and configure available logged email alerts for critical findings.
Limits to check
The Free plan allows 25 hosted scans per month and five per day, with one AI asset under test and a Claude Haiku 4.5 judge; work email verification is required. Team costs $199/mo and allows 300 hosted scans per month, up to 25 scans per day, and a full-strength judge model. The Developer plan costs $799/mo and covers one AI agent or model, one monitored production workflow, monthly automated scans, and single-methodology scheduled scans.
The Club's review
RedLens AI runs automated adversarial assessments against production AI, aimed at teams responsible for securing deployed models and agents. It fits best in regulated organizations that need repeatable testing and audit-oriented reports. Its strongest case is ongoing monitoring tied to security workflows; its main trade-off is that serious production coverage is costly.
Overview
RedLens applies adversarial attacks to production AI and scores findings with CVSS. Its Master Attack Schema v2.0 spans six attack categories and 22 methodologies, giving security teams a structured way to identify vulnerabilities across different attack approaches. Each assessment produces a PDF with a risk score, vulnerabilities, remediation steps, and framework mappings.
Mappings include HIPAA, NIST AI RMF, OWASP GenAI Top 10, MITRE ATLAS, ISO/IEC 42001, FBI CJIS Security Policy, and AIUC-1. They provide evidence teams can use in security work, but they do not certify compliance. That distinction matters for organizations that need to demonstrate controls without treating a scanner report as an attestation.
Key features
Scheduled daily or weekly retests help teams catch drift in production systems, and critical findings can trigger logged email alerts. The cadence is useful for systems that change often, though email is the alert channel available for critical findings; Slack and Teams alerts are not currently available.
A Python CLI and GitHub Action bring scans into CI/CD pipelines, where severity gates can block builds and SARIF output can feed GitHub code scanning. Findings can also stream to Splunk, SentinelOne, Devo, or an HMAC-signed webhook. These options make RedLens more suitable for security teams with established engineering and monitoring workflows than for teams looking for a standalone point-in-time check.
The Blast Radius Mapper accepts agent tool schemas in OpenAPI, Anthropic, and OpenAI formats. That gives teams a way to assess the tools an AI agent can reach, alongside tests of the model or agent itself.
Hosted SaaS is available, while Enterprise adds on-premise Docker deployment and an MSSP console with per-client isolation. The self-hosted package keeps prompts, results, findings, users, and reports in the customer's Postgres database. Attack generation and judging still use Anthropic's API with the customer's key, so on-premise deployment does not remove that external dependency.
RedLens says its 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. Organizations that require RedLens itself to hold that certification should account for the stated timeline.
Pricing
The Free plan costs nothing and includes 25 hosted scans per month, a limit of 5 per day, one AI asset under test, and the Claude Haiku 4.5 judge. Work email verification is required. This is enough to explore the service, but the asset and scan caps constrain ongoing coverage.
Team costs $199 a month, billed $199/mo. It raises the allowance to 300 hosted scans per month and up to 25 per day, removes the cap on assets under test, and includes the full-strength judge model. It suits teams assessing multiple assets with hosted scans, but it does not include the scheduled production monitoring described in the higher plans.
Developer costs $799 a month, billed $799/mo. It covers one AI agent or model, one monitored production workflow, monthly automated scans, and scheduled scans using a single methodology. This is the entry point for recurring production monitoring, but its single workflow and monthly cadence limit its fit for broader or faster-changing deployments.
Professional costs $8,000 a month, billed $8,000/mo. It supports up to 25 monitored production workflows, 50,000 attack simulations per month, weekly automated scans, and unlimited AI assets under test. It is aimed at organizations monitoring a substantial production estate; the monthly price makes it a significant step up from Developer.
Enterprise costs $150,000 a year, billed $150k/yr+. It removes platform caps on workflows, assets, and simulations, adds daily automated scans and self-hosted on-premise Docker deployment, and makes a custom SLA available. This is the plan for large or regulated deployments that need the broadest coverage and deployment control. RedLens offers a free plan and a free trial.
Platforms
RedLens is available through API, self-hosted deployment, and web. The self-hosted option is Enterprise-only and retains an Anthropic API dependency for attack generation and judging.
Who it's for
RedLens is best suited to healthcare, law enforcement, and regulated enterprise organizations deploying patient chatbots, diagnostic AI, internal copilots, or other AI systems. Security teams that want scheduled retesting, severity gates, and framework-mapped reports can connect assessments to existing operational processes. It is a weaker fit for organizations seeking broad employee shadow-AI governance, DLP, model-file scanning, or content moderation, which RedLens does not currently offer.
Pros and cons
- Pros: CVSS-scored findings, remediation steps, and mappings to named frameworks give security teams material they can use in audit and remediation workflows.
- Pros: Scheduled retests, CI/CD build gates, SARIF output, and SIEM streaming support ongoing security operations.
- Pros: Enterprise self-hosting keeps core assessment data in the customer's Postgres database, with per-client isolation in the MSSP console.
- Cons: Developer monitoring starts at $799/mo and covers just one workflow with monthly scans; broader coverage requires a much more expensive plan.
- Cons: Self-hosting still depends on Anthropic's API and the customer's key for attack generation and judging.
- Cons: The product does not currently provide Slack or Teams alerts, broader DLP, employee shadow-AI governance, model-file scanning, or content moderation.
Alternatives
For a wider set of AI red-teaming options, browse AI Red Teaming Tools.
AgentSeal is a freemium option with 30 basic probes on its free plan and a Pro plan that adds 311 probes, runtime MCP scanning with OAuth, and PDF export. Consider it when MCP scanning is central to the choice.
F5 BIG-IP APM is a paid alternative with a free trial and deployment across web, mobile, API, desktop, Linux, macOS, and self-hosted environments. Consider it when those platform options are a priority.
OpenSecureAI Scanner offers a free plan and a $49/month Pro plan with 50,000 API requests per month, hosted AI remediation, and LLM Gateway completions. It may suit readers prioritizing a lower-cost API service with those capabilities.
Promptfoo has a free Community plan with 10k red-team probes per month, all LLM evaluation features, and local or self-hosted operation. Choose it when running evaluations locally or self-hosting on a free plan matters most.
NVADER is another freemium option for readers comparing AI security products.
RedAmon offers a free, MIT-licensed, open-source self-hosted Docker stack for commercial or personal use. It is a fit for readers who prioritize a free self-hosted option.
Rogue offers free personal and internal use, with commercial hosting on a paid plan.
Darkhunt AI Security has a free plan with full platform access and community support, alongside a custom annual plan. It may suit readers looking for a free entry point with access to the full platform.
Verdict
Choose RedLens if your organization needs recurring adversarial assessments, framework-mapped reports, and integrations that connect findings to CI/CD and security operations. Its best fit is a regulated team with budget for production monitoring and a clear use for those workflows. Look elsewhere if you need broad workplace AI governance, want self-hosting without an external attack-generation dependency, or cannot justify the cost of its monitoring plans.
The verdict
RedLens combines adversarial AI assessments with reporting, scheduled retests, and pipeline integrations. Its framework mappings are evidence rather than certification, and self-hosted customers still rely on Anthropic’s API for attack generation and judging.
Questions people ask
Is there a free plan?
Yes. The Free plan costs nothing and includes 25 hosted scans per month, five scans per day, one AI asset under test, and a Claude Haiku 4.5 judge. Work email verification is required.
How much do paid plans cost?
Team is $199/mo, Developer is $799/mo, Professional is $8,000/mo, and Enterprise is $150k/yr+. The listed paid-from price is $799/mo.
Which platforms are available?
RedLens AI is listed for API, self-hosted, and web. Enterprise includes self-hosted on-premise Docker deployment.
Does RedLens AI certify compliance with the listed frameworks?
No. RedLens describes its mappings to HIPAA, NIST AI RMF, OWASP GenAI Top 10, MITRE ATLAS, ISO/IEC 42001, FBI CJIS Security Policy, and AIUC-1 as evidence rather than certification.
What does self-hosted deployment store locally?
The self-hosted package stores prompts, results, findings, users, and reports in the customer’s Postgres. Attack generation and judging still call Anthropic’s API using the customer’s key.
Which integrations does it support?
Findings can stream to Splunk, SentinelOne, Devo, or an HMAC-signed webhook. A Python CLI and GitHub Action support CI/CD scans.
What the maker publishes
- AI agent tools
- Blast Radius Mapper accepts tool schemas in OpenAPI, Anthropic, or OpenAI formats. source
- Assessment report
- An assessment produces a PDF with a risk score, vulnerabilities, remediation steps, and mappings to named frameworks. source
- Attack coverage
- Its Master Attack Schema v2.0 has 6 attack categories and 22 methodologies, with findings scored by CVSS. source
- CI/CD integration
- A Python CLI and GitHub Action can run scans in pipelines, with severity build gates and SARIF output for GitHub code scanning. source
- Deployment
- The site describes hosted SaaS and Enterprise self-hosted on-premise Docker deployment, plus an MSSP console with per-client isolation. source
- Framework mappings
- The site lists mappings to HIPAA, NIST AI RMF, OWASP GenAI Top 10, MITRE ATLAS, ISO/IEC 42001, FBI CJIS Security Policy, and AIUC-1, and says these mappings are evidence rather than certification. source
- Intended users
- The site describes the product for healthcare, law enforcement, and regulated enterprise organizations deploying AI systems. source
- Monitoring
- Scheduled daily or weekly retests can detect drift, while critical findings can trigger logged email alerts. source
- Notable limits
- Slack and Teams alert channels, a Splunkbase app, a PyPI CLI package, broader DLP, employee shadow-AI governance, model-file scanning, and content moderation are listed as not currently available. source
- On-premise data handling
- The self-hosted package stores prompts, results, findings, users, and reports in the customer's Postgres, but attack generation and judging still call Anthropic's API using the customer's key. source
- Purpose
- RedLens runs adversarial attacks against production AI and maps findings to controls used in security audits. source
- Security status
- RedLens says its 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. source
- SIEM integrations
- Findings can stream to Splunk, SentinelOne, Devo, or an HMAC-signed webhook. source
- Support and SLA
- An Enterprise SLA is available, and the Enterprise plan lists a custom SLA as available. source
Checked against the maker's pages on 7 Oct 2026.