Custom Jailbreaks

InjectPrompt offers bespoke jailbreaks tailored to your use case.

Whether you're a developer, researcher, or just curious about what lies beyond the filter, I create and test custom jailbreaks for:

  • OpenAI / ChatGPT

  • Anthropic Claude

  • Google Gemini

  • Custom LLMs

What You Get

  • Bespoke Prompt Chains – Designed for your specific goal (e.g., roleplay, uncensored Q&A, internal logic breaks)

  • Discreet & Secure Delivery – One-to-one via encrypted email or Signal

  • Step-By-Step Reports – Comprehensive documentation detailing methodology and how to replicate


Who Uses This?

  • AI startups testing their models before launch

  • Researchers exploring model boundaries

  • Builders who want full access to their tools


Request a Custom Jailbreak

This is a private, invite-only service. To request access:

  1. Send an email to david@willis-owen.com

  2. Briefly describe your model, your goal, and your risk tolerance.

  3. You’ll receive a quote within 48 hours.

I accept a limited number of jailbreak requests per month to ensure quality and discretion.


Disclaimer

This service is for research, red teaming, and educational purposes only. I do not assist with or condone illegal or unethical use of AI systems.