Group-buying (or volume discount) is a promising field in electronic commerce for applying software agent technologies. In a traditional group-buying mechanism, either a customer or the supplier calls up a sufficient number of buyers for a target item, and then coordinates the actions of all participants during the whole process. Most participants involved in a group-buying project are passive. Studies in this field were therefore focused on developing an effective mechanism so as to enhance the utility of every participant in a fair way. However, the utility of a customer can only be maximized if the customer can buy the item he/she personally needs at a possibly lowest price, not just an item recommended by another customer or the supplier that he/she is supposed to like. In other words, it would be more flexible if every customer can initiate a group-buying project of his/her own for the item he/she personally needs in a convenient way. As a result, there will be multiple group-buying projects for multiple target items at the same time. To this end, a software agent is developed in this study to make every customer easily reach the web page he/she browses for a target item for group-buying. The data of the item will be automatically extracted and uploaded onto a website which then informs every registered user of the group-buying project of this item. Requests for the same item will be combined, and there are always multiple target items on the website for group-buying at the same time. As a result, cross group-buying becomes possible. An experimental system is constructed in this study to demonstrate the applicability of the software agent. Its advantages and/or disadvantages are also discussed.