ncbi-datasets-mcp
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@ncbi-datasets-mcpsearch genome assemblies for Escherichia coli"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
ncbi-datasets-mcp
An MCP server that gives Claude access to NCBI Datasets v2 — search genome assembly metadata, retrieve taxonomy records, and download data packages without leaving your conversation.
Tools
Tool | Transport | Description |
| — | Install the NCBI CLI tools (run once, or set |
| REST | Search genome assemblies by organism name or tax ID |
| REST | Fetch assembly metadata for known accessions |
| CLI | Download a genome package by taxon |
| CLI | Download a genome package by accession |
| CLI | Fetch sequence files for a dehydrated package |
| CLI | Convert a genome JSONL data report to TSV |
| REST | Get lineage, rank, and names for a taxon |
| CLI | Download a taxonomy package |
Related MCP server: mcp-pubmed
Installation
Option 1 — Desktop Extension (recommended for Claude Desktop users)
Download
ncbi-datasets.mcpbfrom the Releases page.Double-click the file and click Install in Claude Desktop.
Optionally enter your NCBI API key and download directory.
The NCBI CLI tools are downloaded automatically on first use (NCBI_AUTO_INSTALL=true is set by default in the extension).
Option 2 — JSON config (Claude Desktop / Claude Code)
Add to claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"ncbi-datasets": {
"command": "uvx",
"args": ["ncbi-datasets-mcp"],
"env": {
"NCBI_API_KEY": "your_key_here",
"NCBI_DOWNLOAD_DIR": "/path/to/downloads",
"NCBI_AUTO_INSTALL": "true"
}
}
}
}Requires uv (curl -LsSf https://astral.sh/uv/install.sh | sh).
Configuration
Variable | Default | Description |
| (none) | NCBI API key — raises rate limit to 10 req/s |
|
| Default download location |
|
| Auto-install CLI tools on startup |
|
| Cap for summary tool result counts |
|
| Seconds before a download times out |
| (auto) | Override path to |
| (auto) | Override path to |
Development
# Install with dev extras
pip install -e ".[dev]"
# Run unit tests
pytest
# Run all tests including live network calls
pytest -m integration
# Regenerate enums from the current NCBI OpenAPI spec
python scripts/gen_enums.py
# Run the server locally (stdio transport)
ncbi-datasets-mcpArchitecture
src/ncbi_datasets_mcp/
server.py FastMCP app — tool registrations only
config.py Pydantic-settings env config
cli/
locator.py Find datasets/dataformat (config → PATH → cache)
installer.py Download binaries from NCBI FTP
runner.py Async subprocess wrapper
rest/
client.py httpx client for metadata/summary endpoints
domains/
_generated_enums.py Vendored enums from OpenAPI spec
common.py Shared utilities (output dir, filename sanitising)
genome.py Genome CLI arg builders + response shaping
taxonomy.py Taxonomy CLI arg builders
models/
responses.py Shared DownloadResult dataclassSummary tools (no file I/O) → REST API.
Download and format-conversion tools → NCBI CLI binaries.
Cite
If you use NCBI Datasets in your research, please cite:
NCBI Datasets. National Center for Biotechnology Information. https://www.ncbi.nlm.nih.gov/datasets/
License
MIT
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/syntheticgio/ncbi-datasets-mcp-server'
If you have feedback or need assistance with the MCP directory API, please join our Discord server