The article describes a computational project that uses browser-based technology to search for new periodic orbits in the three-body problem. The project involves tracing the motion of three masses under Newtonian gravity, testing various initial conditions to find paths that nearly return to their starting points. The system checks these paths at high precision, using a 'bounded Newton shooting' method to refine initial conditions and periods. A 'promising return' is a proposed orbit that is first checked against known orbits in an atlas before being recorded. The system refines these proposals to 106 bits and confirms them in double precision, with some orbits requiring additional verification at higher bit depths. Only verified periodic results are added to the 'checked loops' database, which is labeled as linearly stable or unstable based on small planar perturbations. Known reference orbits are intentionally included in the stream to verify numerical accuracy. The project focuses on orbits that have not been previously cataloged, with the system steering away from families of orbits that have already been extensively documented. The search algorithm prioritizes testing seeds that have previously yielded new orbits or are near already checked loops. The system also tracks the 'braid' of an orbit, which is the sequence of moments when the bodies align, comparing it with the atlas to identify new topologies. Unmatched loops are reviewed to determine if they represent new orbits. The project emphasizes that nothing is automatically added to the atlas; all new orbits must be verified through the described process. The article provides technical details about the computational methods used, including the use of unsoftened Newtonian gravity and the comparison of orbits based on scale-free period and the shape of pairwise distances. The project is designed to be collaborative, leveraging browser-based processing to contribute to the search for new periodic three-body orbits.