Pebblescout is a tool for indexing and searching large nucleotide resources, such as those in the Sequence Read Archive (SRA) and GenBank. It provides indexing and search capabilities to efficiently find relevant subsets of these resources for user queries. Pebblescout uses dense sampling of sequences and ensures match guarantees based on query length. It creates eight databases indexing over 3.7 petabases of sequence data. The tool is available as a web service at https://pebblescout.ncbi.nlm.nih.gov. Pebblescout outperforms existing tools like MetaGraph and Sourmash in terms of accuracy and efficiency for finding relevant subsets of large nucleotide resources. It supports three search modes: Profile, Summary, and Detailed, each providing different levels of detail in search results. Pebblescout is effective for various applications, including finding genes of interest, similar metagenomic runs, and viral sequences with mutations. It provides a data-driven approach to searching large sequence databases, reducing the effort required for downstream analysis. Pebblescout is designed to be efficient and scalable, with a focus on minimizing memory usage and maximizing search performance. It is a general-purpose solution for navigating petabase-scale nucleotide sequence sets, and it is expected to become a standard for future solutions to this problem. The software is not yet production-ready but includes core components and a small example of indexing and search functions. Future work includes improvements to scoring methods, compression schemes, and integration with other analysis tools. The tool is available for public use and encourages user feedback to inform future development.Pebblescout is a tool for indexing and searching large nucleotide resources, such as those in the Sequence Read Archive (SRA) and GenBank. It provides indexing and search capabilities to efficiently find relevant subsets of these resources for user queries. Pebblescout uses dense sampling of sequences and ensures match guarantees based on query length. It creates eight databases indexing over 3.7 petabases of sequence data. The tool is available as a web service at https://pebblescout.ncbi.nlm.nih.gov. Pebblescout outperforms existing tools like MetaGraph and Sourmash in terms of accuracy and efficiency for finding relevant subsets of large nucleotide resources. It supports three search modes: Profile, Summary, and Detailed, each providing different levels of detail in search results. Pebblescout is effective for various applications, including finding genes of interest, similar metagenomic runs, and viral sequences with mutations. It provides a data-driven approach to searching large sequence databases, reducing the effort required for downstream analysis. Pebblescout is designed to be efficient and scalable, with a focus on minimizing memory usage and maximizing search performance. It is a general-purpose solution for navigating petabase-scale nucleotide sequence sets, and it is expected to become a standard for future solutions to this problem. The software is not yet production-ready but includes core components and a small example of indexing and search functions. Future work includes improvements to scoring methods, compression schemes, and integration with other analysis tools. The tool is available for public use and encourages user feedback to inform future development.