TrakEM2 is an open-source software designed for the systematic reconstruction of neuronal circuits from large electron microscopy (EM) and optical image volumes. It addresses challenges in image volume composition, neuronal arbor reconstruction, synapse annotation, and management of large image and annotation collections. The software enables the creation of 3D neural circuits encoded in NeuroML and other formats for analysis. TrakEM2 supports rapid entry, organization, and navigation of tera-scale EM image collections, and offers semi- and automatic image registration across sections. It provides efficient tools for manipulating, visualizing, reconstructing, annotating, and measuring neuronal components. An ontology-controlled tree structure allows hierarchical grouping of reconstructed components based on biologically meaningful entities. TrakEM2 has been successfully used for various applications, including the reconstruction of Drosophila larval central nervous system, array tomography, and the reconstruction of inhibitory networks in mouse visual cortex. The software handles complex imaging arrangements, including low-resolution images of large fields of view and high-resolution images for areas of interest. It supports both linear and non-linear transformations, and allows direct point-and-click manipulation of image transformations. TrakEM2 also facilitates the identification and labeling of neurons and synapses, with manual and semi-automatic methods for image segmentation and sketching. It enables the systematic reconstruction of neuronal circuits, with hierarchical grouping of reconstructed elements. The software is extensible and supports a variety of scripting languages for data manipulation. TrakEM2 is deployed with Fiji, an open-source image processing application, and is available under the General Public License. It has been used in various applications, including the reconstruction of neural circuits in anisotropic serial section EM data. The software is continuously evolving to meet new imaging challenges and is available for modification and incorporation of novel algorithms. TrakEM2 is an essential tool for neuroscience research, enabling the reconstruction of complex neuronal circuits from large image datasets.TrakEM2 is an open-source software designed for the systematic reconstruction of neuronal circuits from large electron microscopy (EM) and optical image volumes. It addresses challenges in image volume composition, neuronal arbor reconstruction, synapse annotation, and management of large image and annotation collections. The software enables the creation of 3D neural circuits encoded in NeuroML and other formats for analysis. TrakEM2 supports rapid entry, organization, and navigation of tera-scale EM image collections, and offers semi- and automatic image registration across sections. It provides efficient tools for manipulating, visualizing, reconstructing, annotating, and measuring neuronal components. An ontology-controlled tree structure allows hierarchical grouping of reconstructed components based on biologically meaningful entities. TrakEM2 has been successfully used for various applications, including the reconstruction of Drosophila larval central nervous system, array tomography, and the reconstruction of inhibitory networks in mouse visual cortex. The software handles complex imaging arrangements, including low-resolution images of large fields of view and high-resolution images for areas of interest. It supports both linear and non-linear transformations, and allows direct point-and-click manipulation of image transformations. TrakEM2 also facilitates the identification and labeling of neurons and synapses, with manual and semi-automatic methods for image segmentation and sketching. It enables the systematic reconstruction of neuronal circuits, with hierarchical grouping of reconstructed elements. The software is extensible and supports a variety of scripting languages for data manipulation. TrakEM2 is deployed with Fiji, an open-source image processing application, and is available under the General Public License. It has been used in various applications, including the reconstruction of neural circuits in anisotropic serial section EM data. The software is continuously evolving to meet new imaging challenges and is available for modification and incorporation of novel algorithms. TrakEM2 is an essential tool for neuroscience research, enabling the reconstruction of complex neuronal circuits from large image datasets.