In the realm of computing, CAPTCHAs are used to determine if a user engaging with a system is a person or a bot. The most common CAPTCHAs are visual in nature, requiring users to recognize images comprising distorted characters or objects. For people with visual impairments, audio CAPTCHAs are accessible alternatives to standard visual CAPTCHAs. Users are required to enter or say the words in an audio-clip when using Audio CAPTCHAs. However, this approach is time-consuming and vulnerable to machine learning algorithms, since automated speech recognition (ASR) systems could eventually understand the content of audio with the improvement of the technique. While adding background noise may deceive ASR systems temporarily, it may cause people to have difficulties de- ciphering the information, thus reducing usability. To address this, we designed a more secure and accessible web CAPTCHA based on the capabilities of people with visually impairments, obviating the need for sight via the use of audio and movement, while also using object detection techniques to enhance the accessibility of the CAPTCHA.