Dryad

What Dryad does

Dryad MCP server gives AI agents access to Dryad, a curated repository of open research datasets. The repository primarily contains data associated with published, peer-reviewed studies across scientific fields. Agents can locate datasets with text searches or retrieve a known record using its DOI.

Search results include identifying and descriptive information such as the DOI, title, authors, a shortened abstract, keywords, publication date, scientific field, and license. A DOI lookup returns more complete metadata and a URL for downloading the dataset files.

How it works

The package exposes two MCP tools: search_datasets and get_dataset. search_datasets accepts a full-text query and searches dataset titles, abstracts, author names, and keywords. get_dataset accepts a Dryad DOI with or without the doi: prefix, such as 10.5061/dryad.hx3ffbgjj.

The hosted connection uses Pipeworx’s MCP gateway at https://gateway.pipeworx.io/dryad/mcp. That scoped gateway endpoint lists the Dryad tools along with shared Pipeworx tools, including routing and discovery functions. The same gateway also offers a full endpoint at https://gateway.pipeworx.io/mcp. Alternatively, the npm package runs as a standalone MCP process and exposes only the Dryad tools.

Setup and configuration

For a hosted MCP connection, add the following server entry to a compatible client:

{
  "mcpServers": {
    "dryad": {
      "url": "https://gateway.pipeworx.io/dryad/mcp"
    }
  }
}

For a local stdio connection, configure the client to run the package with npx:

{
  "mcpServers": {
    "dryad": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-dryad"]
    }
  }
}

The local process communicates through standard input and output. It can also be started directly with npx -y @pipeworx/mcp-dryad to verify that it launches.

The tools can be called without an MCP client through Pipeworx’s HTTP API. For example, a search request can be sent to https://gateway.pipeworx.io/v1/tools/dryad_search_datasets with a JSON body containing a query field.

Tools and capabilities

• search_datasets: Searches Dryad text fields and returns matching dataset summaries.
• get_dataset: Retrieves complete metadata and a file download URL for a DOI.
• DOI input may include either the full doi: form or the bare identifier.
• Results expose publication details, scientific classification, licensing, and dataset version information where available.
• Gateway users can ask ask_pipeworx to route natural-language questions to the appropriate tool.

Limitations and notes

The integration is unofficial and is not affiliated with, endorsed by, or published by Dryad. The hosted endpoint adds shared Pipeworx meta-tools to the scoped connection, so its tool list is larger than the two Dryad-specific tools. The standalone package avoids those gateway-wide tools and provides only the Dryad functionality. Dataset search results contain truncated abstracts, while DOI retrieval provides fuller metadata. Dryad access through this integration is keyless.