
AI-orchestrated continuous pentesting SaaS for web apps and APIs, EU-hosted

AI-orchestrated continuous pentesting SaaS for web apps and APIs, EU-hosted
The Entire Cybersecurity Market, One Prompt Away
Connect your AI assistant to ... tools and ... vendors. Ask anything about the cybersecurity market.
Agent Breach provides continuous offensive security testing for web applications and APIs. The platform is delivered as a SaaS hosted in the EU and does not require agents installed on customer infrastructure. The product runs more than 30 industry-standard testing tools, including Nuclei, SQLMap and Nikto. A large language model (LLM) coordinates these tools: it reads the output of each engine run, selects the next tools via MCP, chains identified weaknesses into attack paths, and explains findings in plain language. The testing itself is performed by deterministic engines, while the LLM handles orchestration and explanation. Key capabilities described by the company: - Findings ranked by exploitability, with remediation explained by an LLM - Authenticated scan coverage - Typical scan duration of about one hour, varying by scope and profile - CI/CD integration and GitHub pull request integration - Authorized scans can be started online The company targets teams that release software more frequently than their traditional pentest schedule allows, and aims to replace periodic pentests with simulation on every deployment. Agent Breach was founded by Denis Cabral Lopes, who leads product, architecture and engineering and is based in Madrid. The website is available in English, Spanish and Brazilian Portuguese.