About GoingLLM
GoingLLM uses GPT-4 to answer complex research questions at the highest level. The AI independently performs web crawls and research, generating answers from simple data, like current temperature, to scientific findings and documents. This revolutionizes efficiency and precision in research. GoingLLM exists as a standalone chatbot, API, and GPT module for integration into ChatGPT. It can be deployed as a Python application on Linux servers or as a Heroku app with one click. Thanks to months of fine-tuning and sophisticated prompting, GoingLLM conducts complex internet searches and crawls with exceptional speed and accuracy, delivering results within seconds.
Founder and Developer
Martin Fürholz is an experienced IT security expert and leading penetration tester from Mecklenburg-Vorpommern, Germany. He has built an AI Red-Teaming practice, focusing on training development, project management, and strategies. He has conducted workshops, including at the University of Rostock for computer science students, and developed practical jailbreak test cases while evaluating various AI systems. Since the first release of ChatGPT, Martin has focused intensively on developing AI bots and tools, especially GoingLLM. He authored a comprehensive book on AI Red-Teaming, serving as internal training material. Beyond these efforts, Martin has made significant contributions to open-source projects and developed several security tools.
Technology
GoingLLM integrates cutting-edge GPT-4 capabilities with a robust backend optimized for rapid, accurate research. Key features include autonomous web crawls, weighted search result evaluation, and validation mechanisms for verifying answers through a secondary research cycle. Deployable as a standalone Python app, Heroku app, API, or GPT module, GoingLLM adapts to various user needs. New functionalities include a vulnerability and exploit search engine, user access controls, chat history, and an advanced multi-user chatbot with internet access.
Project Video
