High-resolution heterogeneous reconstruction of 3D structures of proteins and other biomolecules using cryo-electron microscopy (cryo-EM) is essential for understanding fundamental processes of life. However, it is still challenging to reconstruct the continuous motions of 3D structures from hundreds of thousands of noisy and randomly oriented 2D cryo-EM images. Existing methods based on coordinate-based neural networks show compelling results to model continuous conformations of 3D structures in the Fourier domain, but they suffer from a limited ability to model local flexible regions and lack interpretability. We propose a novel approach, cryoFormer, that utilizes a transformer-based network architecture for continuous heterogeneous cryo-EM reconstruction. We for the first time directly reconstruct continuous conformations of 3D structures using an implicit feature volume in the 3D spatial domain. A novel deformation transformer decoder further improves reconstruction quality and, more importantly, locates and robustly tackles flexible 3D regions caused by conformations. In experiments, our method outperforms current approaches on three public datasets (1 synthetic and 2 experimental) and a new synthetic dataset of PEDV spike protein. The code and new synthetic dataset will be released for better reproducibility of our results. Project page: https://cryoformer.github.io.