RedShield AI
red-shield.ai
- Web
6.9 Club score Visit red-shield.ai
Plans
Every plan and priceEnterprise
—
Quick Scan
$50 one-time
Full Assessment
$100 one-time
Continuous
$250/mo
What the maker publishes
- Adaptive testing
- The testing engine generates deeper follow-up probes when it finds a weakness and moves on when the system blocks an attack category. source
- Adaptive testing
- The testing engine generates targeted follow-up probes when it finds a weakness and moves on from categories the target clearly blocks. source
- Assessment duration
- A Quick Scan typically takes 10 to 15 minutes, while a Full Assessment takes 30 to 60 minutes depending on rate limits and target response times. source
- Assessment duration
- A Quick Scan typically takes 10 to 15 minutes, while a Full Assessment takes 30 to 60 minutes depending on rate limits and target response times. source
- Attack coverage
- The attack library includes prompt extraction, credential probing, jailbreaks, RAG exploitation, tool abuse, indirect prompt injection, hallucination induction, and discriminatory output testing. source
- Attack coverage
- The attack library includes prompt extraction, credential probing, jailbreaks, RAG exploitation, tool abuse, cross-session data probes, hallucination induction, and discriminatory output testing. source
- Configuration
- Customers provide a target URL, define sensitive data patterns and out-of-scope topics, specify the AI’s tools, choose an attack model, and set a rate limit. source
- Internal systems
- The company says it can work with customers on secure access to test systems that are not exposed to the public internet. source
- Live monitoring
- Findings can be monitored in real time during a campaign. source
- Live monitoring
- Users can monitor findings in real time during a campaign. source
- Models
- The company says its attack campaigns use models including Claude by Anthropic and GPT-4 by OpenAI, and that customers can select a model for each engagement. source
- Models
- The maker says its campaigns use models including Claude and GPT-4, and that users can select which model to use for each engagement. source
- Production safety
- The company describes its tests as read-only probes that send prompts and observe responses, and says they do not modify data, execute code, or access backend systems. source
- Production testing
- The maker describes its tests as read-only probes that send prompts and observe responses without modifying data, executing code, or accessing backend systems. source
- Purpose
- RedShield AI runs adaptive security campaigns against AI-powered chatbots and conversational systems, then provides a graded report with evidence and remediation guidance. source
- Purpose
- RedShield AI runs adaptive security campaigns against AI-powered chatbots and conversational systems and provides a graded report with evidence and remediation guidance. source
- Reporting
- Campaign reports include an A through F risk grade, executive summary, detailed findings with exact prompts and responses, remediation guidance, and an attack log appendix. source
- Reports
- Completed campaigns produce a PDF report with an A-through-F risk grade, executive summary, detailed findings with exact prompts and responses, remediation guidance, and an attack log appendix. source
- Setup
- An engagement can be configured with a target URL, sensitive data patterns, out-of-scope topics, AI tools, an attack model, and a rate limit. source
- Support
- Quick Scan includes email support, while Full Assessment and Continuous include priority support. source
- Target systems
- The maker says it can test systems that are not publicly exposed by arranging secure access through its team. source
Checked against the maker's pages on 8 Oct 2026.