llms.txt
Compare with other sites' llms.txt files →# Xint
> Xint (xint.io) is an autonomous application security platform built by Theori. It scans applications through their source code, compiled binaries, or runtime: it maps the attack surface, then runs thousands of AI agents in parallel to find vulnerabilities, test whether they are exploitable, and recommend fixes. The company and the site are called "Xint". Xint Code, Xint Web, Xint Pulse, and Binary Analysis are products within the Xint platform, not separate companies or sites.
Xint is a harness on top of frontier AI models. It makes it cost- and time-effective to secure not just a company's obvious high-value applications but also the "side doors" attackers use to get inside a network: APIs, remote workers, vendors, customer service, and agents.
Xint was built by Theori (theori.io), an offensive security firm founded in 2016 by the team behind 9 DEF CON CTF championships and 150+ disclosed CVEs. The site is available in English (xint.io), Korean (xint.io/ko), and Japanese (xint.io/jp).
## Products
- [Xint Platform](https://xint.io/products/xint-platform): AI penetration testing from code to production. Combines Xint Code (white-box, static testing) and Xint Web (black-box, dynamic testing) to catch vulnerabilities across the full SDLC.
- [Xint Code](https://xint.io/products/xint-platform#xint-code): white-box testing of source code (it can also reverse-engineer compiled binaries), part of the Xint Platform. Reads a codebase the way an attacker would, deciding where to look and following data flows to find complex business logic vulnerabilities, for the most complete view of an application's likely vulnerabilities.
- [Xint Web](https://xint.io/products/xint-platform#xint-web): black-box testing of a running application from its URL alone, the way a real external attacker would, part of the Xint Platform. Maps the most likely attack scenarios and tests them at massive scale, delivering exploitable findings in less than 24 hours on average.
- [Binary Analysis](https://xint.io/products/xint-platform/binary-analysis): one-week analysis of compiled code running on a system (agents, on-premises software, appliances, network daemons, services), with human review and debrief.
- [Xint Pulse](https://xint.io/products/xint-pulse): a single-use, self-directed security test of one web application by the full Xint Web engine, with confirmed, exploitable findings in one report within hours.
- [Public Bug Tracker](https://xint.io/public-bug-tracker): Xint's public security findings in open-source software, discovered autonomously by Xint Code and disclosed responsibly.
## Company
- [About Xint](https://xint.io/about-us): the team behind Xint and Theori, and the credentials behind the platform.
- [Contact](https://xint.io/contact-us): request a pilot, live demo, or technical conversation.
## Resources
- [Resource Center](https://xint.io/resources): research, findings, technical write-ups, webinars, and whitepapers from Xint's offensive security team.
- [IDC Innovators: Autonomous Penetration Testing for DevSecOps, 2026 (excerpt)](https://xint.io/resources/idc-innovators-autonomous-penetration-testing-for-devsecops-2026-excerpt): whitepaper.
- [The Top Security Vulnerabilities Generated by AI Code](https://xint.io/resources/the-top-security-vulnerabilities-generated-by-ai-code): whitepaper.
- [ARTEX Case Study and Defense Guide](https://xint.io/resources/artex-case-study-and-defense-guide): whitepaper.
- [Xint Reinvents AppSec Testing for the LLM Era](https://xint.io/resources/xint-reinvents-appsec-testing-for-the-llm-era): video.
- [AI-Assisted AppSec: How to Validate and Patch Vulnerabilities Faster Without Losing Control](https://xint.io/resources/should-you-allow-fully-autonomous-ai-remediation): webinar.
- [The Side Door Problem: What the Korean Bank Breaches Mean for US Banks](https://xint.io/resources/the-side-door-problem-what-the-korean-bank-breaches-mean-for-us-banks): webinar.
- [What Does a Bug Actually Cost?](https://xint.io/resources/what-does-a-bug-actually-cost): webinar.
- [What Security Flaws Is AI Prone to Create](https://xint.io/resources/what-security-flaws-is-ai-prone-to-create): webinar.
- [Ultimate Web Hacking | Black Hat Briefing Series](https://xint.io/resources/ultimate-web-hacking-blackhat-briefing-series): webinar.
## Blog
- [Blog](https://xint.io/blog): research, vulnerability write-ups, and company news from the Xint team.
- [Kia's Red Team Fixes 50% More Vulnerabilities With the Same Headcount Using Xint](https://xint.io/blog/kia-red-team-xint): customer story.
- [Xint Selected for Anthropic and OpenAI's Cyber Programs](https://xint.io/blog/xint-selected-for-anthropic-and-openais-cyber-programs-377371): company news.
- [DARPA Selects Xint to Research Source Code Vulnerabilities in Critical Messaging Applications](https://xint.io/blog/darpa-selects-xint-to-research-source-code-vulnerabilities-in-critical-messaging-applications-used-in-military-communications-330339): company news.
- [WordPress Core Update Race: When the Update Becomes the Attack Path](https://xint.io/blog/wordpress-core-update-attack): research.
- [Stealing OpenAI Reasoning Traces via Compaction Smuggling](https://xint.io/blog/compaction-smuggling-reasoning-traces): research.
- [PHP2Shell (CVE-2026-109618/109619) Emergency Response Plan](https://xint.io/blog/php2shell-emergency-response): research.
- [It Wasn't the AI Model That Was Breached: 5 Recent AI Agent Vulnerabilities and Incidents](https://xint.io/blog/ai-agent-security-vulnerabilities): research.
- [Any Android App Can Shut Down the Phone: Found in Three Hours Without Source Code](https://xint.io/blog/android-app-shutdown-vulnerability): research.
## Other languages
- [Xint (Korean)](https://xint.io/ko): Korean version of the site.
- [Xint Blog (Korean)](https://xint.io/ko/blog): Korean blog.
- [Xint (Japanese)](https://xint.io/jp): Japanese version of the site.
## Optional
- [Privacy Policy](https://xint.io/privacy)
- [Terms of Use](https://xint.io/terms)