I Built an Tool to AI Agent Testing

The author introduces the AI Agent Tester, an open-source project designed to automate the validation of AI model responses. Manual prompt testing is inefficient, motivating the creation of this tool. It reads prompts from a CSV, sends them to an AI model, and checks the responses for specific keywords. The tester employs stemming via NLTK to identify variations of expected words, enhancing validation accuracy. A JSON report detailing the success or failure of each prompt, along with the AI's complete response, is generated. The tool is built in Python, ensuring ease of use and minimal setup. Automatic proxy support is included for use in corporate environments. This project is targeted towards developers and QA engineers who integrate large language models. The author encourages feedback and contributions to the open-source project, hosted on GitHub.