The AI pentesting agent for your terminal

RedCell is for systems you own or have written permission to test. Define the authorized targets and rules of engagement before testing. Testing systems without permission may be illegal and violates RedCell’s Acceptable Use Policy.

RedCell runs on macOS (Apple Silicon), Linux, and Windows. Install it from a terminal on macOS or Linux, or use the PowerShell installer on Windows. The agent runs on your machine.

RedCell plans start at $20 per month: Starter $20, Pro $99 and Max $199 per month, or $16, $79 and $159 per month billed yearly. Creating an account is free, but CLI inference needs a paid plan. New accounts can try any monthly plan for 24 hours for $1.

AI penetration testing uses AI models or agents to assist with parts of an authorized security assessment, such as reconnaissance, scanning, vulnerability validation, and reporting. A human operator defines scope and reviews results.

AI can speed up repetitive work such as reconnaissance and CVE checks, but human testers still provide scope decisions, judgment, review, and accountability. RedCell is designed to support an operator, not remove that responsibility.

Metasploit and nuclei provide security testing frameworks and scanners. RedCell is a terminal-based AI agent that can plan and coordinate work across tools, preserve engagement context, and organize findings with evidence. RedCell can use nuclei templates in its workflow.

RedCell applies scope guardrails to help keep activity within the targets and rules of engagement you define. Its security policy blocks out-of-scope traffic, rate-limit violations, and prohibited high-impact actions.

RedCell can help automate passive OSINT, subdomain enumeration, port and service fingerprinting, template-based CVE scanning, scoped vulnerability verification, evidence collection, and Markdown or JSON report export.

Yes. RedCell can connect to custom MCP servers, Python scripts, and native security binaries. See the CLI documentation for setup details.

RedCell is a paid terminal product with model usage included in paid plan budgets, integrated scope guardrails, and structured usage and billing in one service. Check the pricing page and product documentation for current plan features.

RedCell runs on your machine, but model turns needed for reasoning are sent to its inference backend. RedCell says provider data retention is set to deny and that it does not store prompt or file content beyond lightweight session aggregates needed to run and resume work. Read the Security and Privacy pages for details.