Skip to content
Try Today
Try Today
QUILT PACKAGES:

THE NEW STANDARD FOR DATA ASSETS

Begin building Quilt Packages today with Open Source Software
Asset 9
QUILT OPEN SOURCE GETS YOU STARTED
  • Store contextual metadata alongside the data so teams can find the exact data they're looking for.
  • Track each version of the data so teams know they're using the same version every time.
  • Link data to code so teams can run the same analysis every time.
MEET WITH OUR TEAM

Powerful
Open Source
AI-Ready Data Management

Build once. Reuse everywhere.
for people and AI.

Quilt’s open source tools build versioned, context-rich data packages that are secure, governed, and AI-ready. Every package becomes a trusted foundation for scientists and large language models alike.

Get Started
View on GitHub
Terminal
$ pip install quilt3
$ quilt3 install username/dataset
$ quilt3 package show quilts/open-data
│ ├─ README.md
│ ├─ data/
│ │ ├─ samples.csv
│ │ └─ metadata.json
✓ Package installed successfully
Open Source

Everything you need to build packages

Quilt's open-source tools turn data in your own S3 into versioned, AI-ready packages, from your first pip install to a fully governed catalog.

quilt3 Python client

Build, push, and version data packages straight from a notebook or script. pip install quilt3 and your S3 data becomes reproducible.

$ pip install quilt3

Open-source Quilt catalog

A self-hosted web catalog to browse, search, and visualize every package in your own S3 buckets. The data stays inside your AWS account.

  • File tree, README & metadata in one view
  • In-browser previews for images & tables

Versioned packages

Every package is immutable and addressed by a content hash, so it stays reproducible, tamper-evident, and audit-ready for GxP and 21 CFR Part 11.

$ quilt3 push team/dataset

README & quilt_summarize.json

Document datasets in Markdown and configure rich previews with quilt_summarize.json. Charts, galleries, and tables render automatically in the catalog.

  • Self-describing, context-rich packages
  • AI-ready: agents read the same context
Built for AI

Packages are where AI in life sciences starts

A model is only as good as the data it can find and trust. A Quilt package bundles your data with its metadata and a versioned, cryptographic history. That gives models and agents the context they need, and lets you trace any result back to the exact data that produced it.

Packages you can verify

Data, metadata, and an immutable version history in one addressable unit. Every output traces back to an exact dataset, so it stays reproducible and audit-ready for regulated work.

Qurator: search in plain English

Ask for data the way you'd ask a colleague. Qurator searches your governed catalog and returns the right packages, scoped to what each person is allowed to see. No ontology expertise needed.

Bring your own model (MCP)

Connect Claude, ChatGPT, or any MCP-compatible agent to your Quilt data with per-user OAuth. Models can read, visualize, and build on your data, and they only ever see what that user can.

From instrument to model, your data stays in your AWS account: governed, versioned, and ready to use.

GET STARTED TODAY

To get started building Quilt Packages today, take a look at our Quickstart or visit our Github Repo

GitHub Go to the GitHub Repo
Visit the Quickstart

Trusted by life-sciences organizations

Data lookups that used to take our scientists days now take minutes, with a single, governed source of truth the whole team can trust.
90% faster data lookup Data Platform team, Resilience