# Dataverse Harvard

> Dataverse Harvard is a freemium MCP server for academic research. MCP server for searching and retrieving Harvard Dataverse research datasets and files. It can run locally on your machine or as a hosted remote server.

- **Kind:** MCP server
- **Pricing:** Freemium
- **Starting price:** $49 a month
- **Platforms:** API, CLI, self-hosted
- **Website:** https://pipeworx.io/packs/dataverse-harvard
- **Repository:** https://github.com/pipeworx-io/mcp-dataverse-harvard
- **MCP Registry name:** io.github.pipeworx-io/dataverse-harvard
- **Categories:** [academic research](https://ai-powered.com/categories/academic-research)
- **Listing updated:** 2026-11-18

## About

The server gives AI clients read access to Harvard Dataverse, an open research data repository with roughly 150,000 to 200,000 datasets depending on the source. Keyword search returns dataset titles, DOIs, authors and publication dates, and a DOI taken from a search result can then be passed to tools that return full metadata or list a dataset's files with download URLs.

The connection is served through the Pipeworx gateway, either as a hosted streamable HTTP endpoint or as an npm stdio bridge that forwards calls to it. The README notes that a pack endpoint also lists shared gateway meta-tools such as ask_pipeworx alongside the four dataset tools, and that the npm bridge lists only the pack's own tools unless an environment flag is set.

The server suits researchers, data librarians and developers building agents that need to find published datasets by topic and check their contents. Reads are keyless, though the README says anonymous calls are capped daily and a free Pipeworx key raises the limit.

## Features

- **Keyword dataset search:** Searches Harvard Dataverse for datasets, files or dataverses with optional type, sort and paging parameters.
- **Dataset metadata lookup:** Returns full metadata for one dataset by its DOI-style persistent identifier.
- **Dataset file listing:** Lists files in the latest version of a dataset with names, content types, sizes and download URLs.
- **Dataverse collection metadata:** Retrieves metadata for a dataverse collection by alias or id.
- **Keyless read access:** Reads do not require an account, though anonymous use is rate limited per day.

## Good for

- Finding published research datasets on a given topic for a literature or data review
- Checking which files a Dataverse dataset contains before downloading it
- Retrieving authors, abstracts and publication dates for citing a dataset
- Letting an AI agent answer questions about Harvard-hosted research data

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http dataverse-harvard https://gateway.pipeworx.io/dataverse-harvard/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://gateway.pipeworx.io/dataverse-harvard/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "dataverse-harvard": {
      "url": "https://gateway.pipeworx.io/dataverse-harvard/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "dataverse-harvard": {
      "type": "http",
      "url": "https://gateway.pipeworx.io/dataverse-harvard/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "dataverse-harvard": {
      "serverUrl": "https://gateway.pipeworx.io/dataverse-harvard/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "dataverse-harvard": {
      "type": "http",
      "url": "https://gateway.pipeworx.io/dataverse-harvard/mcp"
    }
  }
}
```

## Tools

- `search`: Keyword search across datasets, files and dataverses.
- `dataset`: Full metadata for one dataset by persistent identifier.
- `dataset_files`: Lists files in a dataset's latest version.
- `dataverse`: Metadata for a dataverse collection by alias or id.

## FAQ

### Does it need an API key?

The README says read calls are keyless. Without a Pipeworx key, calls are anonymous and capped daily; a free key raises the cap.

### How do I get a dataset's DOI for the other tools?

The README says the dataset and file tools need a persistent identifier, which should come from a search result rather than being guessed.

### Which upstream API does it use?

The README names the Harvard Dataverse API at dataverse.harvard.edu/api as its data source.

### Is Dataverse Harvard free?

Dataverse Harvard has a free plan, with paid plans for more usage or features. Paid plans start at $49 a month.

### How do I add Dataverse Harvard to Claude Code?

Run this command in your terminal: claude mcp add --transport http dataverse-harvard https://gateway.pipeworx.io/dataverse-harvard/mcp

### What are the best alternatives to Dataverse Harvard?

The closest alternatives to Dataverse Harvard are Dryad, Figshare, Zenodo, Semanticscholar and Paperswithcode.

## Where this comes from

- Description: written from the project's README, the project's website and its MCP Registry entry.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [Dryad](https://ai-powered.com/mcp/dryad): Dryad is a freemium MCP server for academic research.
- [Figshare](https://ai-powered.com/mcp/figshare): Figshare is a freemium MCP server for academic research.
- [Zenodo](https://ai-powered.com/mcp/zenodo): Zenodo is a freemium MCP server for academic research.
- [Semanticscholar](https://ai-powered.com/mcp/semanticscholar-2): Semanticscholar is a freemium MCP server for academic research.
- [Paperswithcode](https://ai-powered.com/mcp/paperswithcode): Paperswithcode is a freemium MCP server for academic research.

---
Source: https://ai-powered.com/mcp/dataverse-harvard
