Binning is an essential step in genome-resolved metagenomic analysis in which assembled contigs originating from the same source population are clustered. However it is challenging, especially for low abundance microbial species. Here the authors introduce a toolkit that integrates multiple prominent binning tools and AI for efficient and high-resolution recovery of non-redundant bins from short- and long-read metagenomic sequencing datasets.