AI Guide
Artificial Intelligence (AI) is a growing component of the Dataverse ecosystem.
Tools
Ask Dataverse
Ask Dataverse (ask.dataverse.org) is a place to ask questions about the Dataverse Project and the Dataverse software. It was created by Slava Tykhonov who announced it in December 2024 and presented it February 2025 (video, slides, notes).
Ask the Data
Ask the Data is an external tool that allows you ask natural language questions about the data contained in Dataverse tables (tabular data). See the README.md file at https://github.com/IQSS/askdataverse/tree/main/askthedata for the instructions on adding Ask the Data to your Dataverse installation.
AutoSage
AutoSage provides metadata suggestions for datasets. Learn more at https://github.com/siacus/autosage.
DataChat
DataChat is a multilingual open source natural language interface for Dataverse and other data platforms with an experimental Graph AI implementation for Croissant support. DataChat can literally talk back to you and explain what is inside of every single dataset, you can ask any question and it responds on the level of metadata described by Croissant standard. Learn more at https://github.com/gdcc/datachat.
TurboCurator
TurboCurator is an external tool that generates metadata improvements for title, description, and keywords. It relies on OpenAI’s ChatGPT & ICPSR best practices. See the TurboCurator Dataverse Administrator page for more details on how it works and adding TurboCurator to your Dataverse installation.
Protocols
Model Context Protocol (MCP)
Model Context Protocol (MCP) is a standard for AI Agents to communicate with tools and services, announced in November 2024.
An MCP server for Dataverse has been deployed to mcp.dataverse.org, powered by the code at https://github.com/gdcc/mcp-dataverse. See the code’s README for information on configuring MCP clients (e.g. Cursor, Visual Studio Code, Windsurf, Zed, etc.) to use mcp.dataverse.org or your own local installation (setup instructions are also provided).