The detection of hazardous gases, such as carbon monoxide (CO), is crucial for ensuring human safety and health, especially in enclosed environments or areas with a high potential for exposure to toxic gases. Carbon monoxide is a colorless, odorless gas that is extremely dangerous when inhaled in large amounts. Therefore, a CO detection system that can provide early warnings is essential. This research aims to design a carbon monoxide (CO) detector using the MQ-7 sensor, which can detect CO concentrations in the air. The detection system is developed based on Internet of Things (IoT) technology, enabling real-time CO detection data to be transmitted to a cloud platform or user devices via an internet connection. This allows users to monitor CO levels in their environment anytime and anywhere. The system consists of the MQ-7 sensor to detect CO levels, a microcontroller as the controller, and a communication module to send data to the internet. The main advantage of this system is its ability to notify users if CO levels exceed a safe threshold. Thus, this detector can function as a preventive tool against CO exposure accidents. The use of IoT in this system also facilitates remote monitoring, improving the effectiveness and efficiency of managing a safe environment.

Most door accesses used in homes still use conventional keys. Where the use of manual keys is still vulnerable to loss or damage and is not very safe to use, because manual keys can be duplicated or duplicated by other people. This certainly feels less practical and flexible. This research aims to help improve door security. With a voice recognition system, only certain people can access it using a certain access code. Researchers used the trial and error method. The factor being tried is the voice recognition V3 module which can read voice recordings. The results of this research can be implemented in the form of a tool, where the system can open and close house doors according to the planned design, namely through voice commands and the designed system can run normally according to the automatic system that has been designed in prototype form. The conclusion from this research is that the voice recognition module is capable of recording sounds up to a distance of 20 cm. In voice recording < 10 cm the module is relatively easier to receive voice commands, in recording > 10 cm and < 15 cm the module is slightly less able to receive voice commands, so that in voice recording at a distance of 15 cm to 20 cm the module is very poor in receiving commands sound well.