Using & Downloading Datasets

Every dataset has a detail page at /datasets/<slug> with the dataset card, the file browser, and the clone panel.

Read the dataset card

The card (rendered from the repo's README.md) tells you what's in the dataset: fields, splits, size, collection method, license and limitations. Check the license before downloading.

Browse & preview files

The file browser lists everything in the dataset's object storage — Parquet, CSV, WebDataset shards, media. You can preview small text files in the in-browser code editor and download any file directly.

Download

In the browser

Open the dataset page, find the file, and click download.

With curl

Each file resolves through a stable URL that 302-redirects to a presigned / CDN link:

# $API = https://api.chaimind.in (or your local backend)
curl -L "$API/api/v1/datasets/<slug>/resolve/data/train.parquet" -o train.parquet

-L follows the redirect to the bytes. Public datasets need no auth; private ones require -H "Authorization: Bearer <token>".

List files via the API

curl "$API/api/v1/datasets/<slug>/files"

Returns the file list (paths + sizes) backing the browser.

Clone the code repo

Loading scripts and docs live in an auto-provisioned git repo, separate from the data files. Copy the URL from the clone panel:

git clone <clone-url>

Two stores, by design. Text/code → git (clone it). Data files → object storage (download via resolve). See Publishing Datasets for the full split.

Next: Browsing & Discovery · Publishing Datasets · Import from GitHub.