CyCraft Launches XecGuard Community Support Program! Applications Open Now for AI Developers and Open-Source Communities

As AI Agent applications are deployed at an unprecedented pace, the corresponding security mechanisms remain critically insufficient. To address this, CyCraft is launching the XecGuard Community Support Program to empower the AI developer community in building a more secure AI ecosystem. Beyond providing tangible resources, this program advocates for a core principle: "AI Safety is a standard, not an option." Only by integrating security into the core architecture from the design phase can AI systems achieve long-term trust and scalable development.

Qualified open-source developers and contributors can receive a free three-month subscription to XecGuard Lite by completing the application form. Through this program, CyCraft aims to lower the barrier for developers to implement AI security, enabling projects to maintain safety and stability alongside rapid innovation.

【How to Apply】 Please fill out the application form on the XecGuard Community Support Program website: https://go.cycraft.ai/xecguard-csp

‍

XecGuard Community Support Program

※ Please read the full program description and terms of use carefully before applying. CyCraft reserves the right of final review and approval for all applications, as well as the right to modify, change, or terminate this program at any time.

1. Eligibility: Applicants must use a GitHub account; the associated open-source project must have at least 100 Stars.
2. Non-Commercial Use Only: XecGuard Lite is designed for individuals and open-source communities. It may not be used for any commercial scenarios, profit-making activities, or purposes that violate local laws and regulations. CyCraft reserves the right to terminate XecGuard services for any violations of these terms.
3. Technical Specifications: XecGuard Lite is optimized for security detection in chat and short-form text content. API limits are set to a maximum of 180 requests per minute (RPM), supporting a 128K Context Window (analysis is performed on the last 10K Tokens of plaintext).
4. Use Scenario: Best suited for real-time human interaction scenarios. It does not support scanning of full files, audio/video, images, or excessively long data sets.
5. Data Privacy: Content from normal requests will not be logged or retained. Only when a malicious attack (such as a Prompt Attack) is detected will the relevant content be de-identified for use in security research, model optimization, and AI safety advocacy.

In addition to resource support, CyCraft is launching a dedicated Discord community. Whether you are a veteran AI developer or an enthusiast interested in AI security and Prompt Attack trends, we welcome you to join the discussion and exchange insights with CyCraft’s technical experts.

Learn more about XecGuard AI Guardrail or Contact Us.