Data is better when we work together.

qri (“query”) is versioned, scriptable, exportable, collaborative datasets

A Web of Datasets

Qri is built around datasets. Bigger than a spreadsheet, smaller than a database, datasets are all around us. Use Qri to browse, download, create, fork, and publish datasets with a broad network of peers.

This Party is Free and Open Source

Data is better when we work together. Qri costs nothing to use, and is built as an open source project under a GPL license.

Datasets You Can Actually Use

Every dataset change is tracked & attributed to an author, so you can audit whether the data you’re looking at meets your standards, and track changes as they happen.

Tools for Any Skill Level

Whether you're a data scientist, or have only ever touched excel, we have tools for you.

Built on the Distributed Web

Qri is built from the ground up as a distributed network on top of IPFS. We chose IPFS because it’s both global and content-addressed — perfect for datasets.

Data you’ve downloaded stays local. Content-addressing lets data be stored anywhere without sacrificing security. All this adds up to a web of datasets that is faster, more secure, and free.

Works With Both
Mouse and Keyboard

$ qri add --file=dataset.yaml me/datadataset created!
$ qri connectconnecting to IPFS and qri P2P…
peername: b5
JSON API port: 2503
Webapp port: 2505

Qri has a desktop app and command line tools. Both are free and open source.

Qri Uses Existing Specs

Wherever possible, we aim to use specifications & technologies that already exist. The end result is a natural set of integration points that makes qri less about being a “data platform” and more a series of integrations between platforms.

Git-style version control

Qri’s dataset versioning system is inspired by git, and signs each commit with your identifying keypair. Because qri is only about datasets, qri generates commit messages for you.

Native support for JSON, CSV, CBOR data formats

Mix & match any format as you need, import from and export to any format.

Metadata based on library science

Librarians are better at metadata than developers, so we based our metadata spec on DCAT & Project Open Data, for cleaner integration with existing data catalogs.

JSON-Schemas for validation & OpenAPIs

Dataset schemas are defined with the same spec that drives OpenAPIs. Datasets automatically generate a JSON API & accompanying OpenAPI documentation.

Automate data munging with Python’s cousin: Starlark

Write configurable, repeatable transformations that can build on remote sources and other qri datasets, in a syntax that feels like Python.

Data is Better When We Work Together

Join our mailing list!

At-most-once-a-month updates on project progress.