This dataset contains three dimensional STL dental surface meshes obtained from intraoral scans of patients undergoing fixed prosthodontic treatment. Each scan represents a cleaned and standardized surface model of the dental arch and is paired with structured annotation and metadata files. For every STL file, a corresponding JSON file provides spatial annotations for all prosthetic objects present in the scan. These annotations include axis aligned three dimensional bounding boxes defined by centroid coordinates, bounding box dimensions along the X, Y, and Z axes, and the minimum and maximum coordinate values describing object extent. All spatial values are reported in millimeters and refer to a common coordinate system established through principal component analysis based standardization. In addition, a tabular metadata file summarizes all annotated restorations and links each object to its source scan. The metadata include the scan identifier, restoration identifier, class label, tooth type, arch quadrant, and tooth number. This structure enables systematic filtering, stratified analysis, and reproducible evaluation across anatomical regions and restoration categories. Together, the STL geometry files, JSON annotations, and metadata table provide a coherent and interoperable dataset for three dimensional object detection, mesh based analysis, and geometric deep learning research in digital dentistry.