ncbi-datasets-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| NCBI_API_KEY | No | NCBI API key — raises rate limit to 10 req/s | |
| NCBI_CLI_PATH | No | Override path to datasets binary | |
| NCBI_MAX_RESULTS | No | Cap for summary tool result counts | 20 |
| NCBI_AUTO_INSTALL | No | Auto-install CLI tools on startup | false |
| NCBI_DOWNLOAD_DIR | No | Default download location | ~/Downloads/ncbi_datasets |
| NCBI_DATAFORMAT_PATH | No | Override path to dataformat binary | |
| NCBI_REQUEST_TIMEOUT | No | Seconds before a download times out | 300 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| ensure_cliA | Check whether the NCBI datasets/dataformat CLI tools are installed and, if not, download and install them automatically. Call this tool once before using any download or format-conversion tools. Returns the install status, binary paths, and version string. |
| list_data_typesA | Describe the kinds of data NCBI Datasets can provide. Call with no arguments for a readable overview of every NCBI Datasets data report type (genes, genome assemblies, sequences, taxonomy, viruses, and more), including which other tools in this server retrieve each one. Pass a report_type (e.g. "genome-assembly", "taxonomy", "virus") to get that type's full field categories and a link to its schema documentation. Args: report_type: Optional data report type key. Omit to list all types. Unknown values return the list of valid keys. |
| genome_summary_by_taxonA | Search NCBI for genome assemblies matching a taxon name or tax ID. Returns assembly metadata including accession numbers, organism info, assembly statistics, annotation status, and submission details. Args: taxon: Taxon name (e.g. "human", "Mus musculus") or NCBI tax ID assembly_level: Filter by level — chromosome, complete_genome, contig, scaffold assembly_source: Filter by source — all, genbank, refseq reference_only: Return only reference/representative genomes annotated_only: Return only annotated genomes limit: Maximum number of assemblies to return (capped by server config) |
| genome_summary_by_accessionA | Retrieve genome assembly metadata for one or more NCBI accessions. Args: accessions: RefSeq or GenBank assembly accessions (e.g. ["GCF_000001405.40", "GCA_000001405.29"]) |
| genome_download_by_accessionA | Download a genome data package for one or more assembly accessions. Returns the local path to the downloaded ZIP file. Args: accessions: RefSeq/GenBank assembly accessions include: Data types — genome, rna, protein, cds, gff3, gbff, seq-report. Defaults to ["genome"]. output_dir: Directory for the download. Uses NCBI_DOWNLOAD_DIR if omitted. dehydrated: Download a lightweight dehydrated package (use rehydrate_genome_package to fetch the actual sequence files). |
| genome_download_by_taxonA | Download genome data packages for assemblies matching a taxon. For taxa with many assemblies, use dehydrated=True first, then call rehydrate_genome_package to fetch only the data you need. Args: taxon: Taxon name or NCBI tax ID include: Data types — genome, rna, protein, cds, gff3, gbff, seq-report output_dir: Directory for the download dehydrated: Download a dehydrated package (recommended for large taxa) reference_only: Download only the reference/representative genome |
| rehydrate_genome_packageA | Fetch the full sequence files for a dehydrated genome package. Args: package_path: Path to the dehydrated ZIP file or its extracted directory. |
| dataformat_genome_tsvB | Convert a genome JSONL data report from a downloaded package to TSV. Args: jsonl_path: Path to the genome data_report.jsonl file inside the package. fields: Optional list of fields to include. Omit to use the default set. |
| taxonomy_summaryA | Get taxonomy metadata for a taxon — lineage, rank, scientific name, common names, and child taxa. Args: taxon: Taxon name (e.g. "human", "Bacteria") or NCBI tax ID |
| taxonomy_downloadC | Download a taxonomy data package for a given taxon. Args: taxon: Taxon name or NCBI tax ID output_dir: Directory for the download |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 10 tools
Each tool has a clearly distinct purpose: setup, summary/download by accession or taxon, taxonomy, format conversion, and data type listing. No overlap in functionality.
Most tools follow a consistent resource_action_by_criterion pattern (genome_download_by_accession, genome_summary_by_taxon), with minor deviations like dataformat_genome_tsv and ensure_cli. Still highly readable.
10 tools is well-scoped for NCBI Datasets: provides setup, summary, download, conversion, and overview without being overwhelming.
Covers the full lifecycle: setup, search/summary, download (including dehydrated/rehydration), format conversion, and data type listing. No obvious gaps for the domain.