296 datasets · 25 topics ·20 hot for AI

The world's data.
Delivered.

From GDP to Common Crawl, ImageNet to Sentinel-2. The most popular and most in-demand datasets, with licence terms up front, delivered to your inbox, Teams, Slack, spreadsheet, database or bucket.

Private beta: accounts are opening to a small group first.

Popular datasets
296
Values on record
1.91M
Series to choose from
27.6K
History back to
1 AD

Hot for AI.

The 20 datasets AI start-ups want most right now, with their licence terms up front.

See all 20
Via publisher

ImageNet

ImageNet (Stanford and Princeton)

14 million images; ILSVRC subset ~150 GB

Coming soon

COCO (Common Objects in Context)

COCO Consortium

330,000 images, ~25 GB

Coming soon

Common Crawl

Common Crawl Foundation

Petabytes; ~2-3 billion pages per crawl

Coming soon

OpenStreetMap

OpenStreetMap Foundation

Planet file ~80 GB (PBF)

Coming soon

Wikipedia (full dumps)

Wikimedia Foundation

300+ languages; English articles ~25 GB compressed

Via publisher

Re-LAION-5B

LAION e.V.

5.5 billion image links with captions

Via publisher

Amazon Reviews

McAuley Lab, UC San Diego (Amazon Reviews 2023)

571 million reviews, 48 million products

Coming soon

SQuAD (Stanford Question Answering)

Stanford NLP

150,000 questions

Via publisher

MIMIC-IV

MIT Laboratory for Computational Physiology (PhysioNet)

300,000+ patients

Coming soon

FineWeb

Hugging Face

15 trillion tokens

Coming soon

GitHub public code

GitHub (via GH Archive and public repositories)

Hundreds of millions of public repositories

Coming soon

Open Images

Google

9 million images

Via publisher

Waymo Open Dataset

Waymo

Thousands of driving scenes

Via publisher

SA-1B (Segment Anything)

Meta AI

11 million images, 1.1 billion masks

Coming soon

Common Voice

Mozilla Foundation

30,000+ hours in 130+ languages

Via publisher

LMSYS Chatbot Arena conversations

LMSYS

1 million conversations with 25 models

Coming soon

UltraChat

Tsinghua University (THUNLP)

1.5 million dialogues

Coming soon

Google Open Buildings

Google Research

1.8 billion building footprints

Coming soon

Sentinel-2

European Space Agency (Copernicus)

Petabytes; the whole land surface every 5 days

Coming soon

ERA5 climate reanalysis

Copernicus Climate Change Service

Every topic that matters.

Not just finance. Whatever the world measures and AI learns from, on a schedule.

All datasets

AI training text

13

Common Crawl · Wikipedia (full dumps) · FineWeb

Images and video

24

ImageNet · COCO (Common Objects in Context) · Re-LAION-5B

Language and AI benchmarks

13

SQuAD (Stanford Question Answering) · LMSYS Chatbot Arena conversations · UltraChat

Code

5

GitHub public code · CodeSearchNet · StarCoderData

Audio and speech

9

LibriSpeech · Common Voice · AudioSet

Economy and growth

24

GDP · Inflation (consumer prices) · GDP per person

Population

15

Population · Fertility rate · Population, long run (10,000 BC to today)

Climate and environment

18

CO₂ emissions per person · Global temperature change · Annual CO₂ emissions

Satellite and Earth observation

4

Sentinel-2 · Sentinel-1 · Landsat

Energy

23

Electricity from renewables · Energy use per person · Electricity from solar

Health

21

Life expectancy at birth · Life expectancy, long run · Child mortality, long run

Biology and medicine

10

MIMIC-IV · PhysioNet · The Cancer Genome Atlas (TCGA)

Markets and finance

16

NY Fed reference rates (SOFR, EFFR, OBFR) · US Treasury par yield curve · Foreign exchange rates (H.10)

Jobs and labour

9

Unemployment rate · Youth unemployment · Labour force participation

Trade and investment

8

Exports · Imports · Foreign direct investment (inflows)

Society and inequality

13

Titanic passengers (Kaggle) · Extreme poverty ($3 a day) · Income inequality (Gini index)

Government

7

Government debt · Tax revenue · Military spending

Education

7

Education spending · Average years of schooling · Adult literacy

Technology and internet

10

Internet users · Mobile phone subscriptions · Number of internet users

Food and agriculture

12

Undernourishment · Daily calories per person · Meat production

Transport and travel

9

Electric car sales share · International tourist arrivals · Electric car sales

Science and space

5

Iris flowers · Research and development spending · Patent applications

Companies and business

9

Amazon Reviews · Open Security & Entity Master · SEC EDGAR company filings

Places and geography

5

OpenStreetMap · Google Open Buildings · Population density

Weather and hazards

7

Deaths from natural disasters · Earthquakes worldwide · Daily weather observations (GHCN)

The most popular data in the world.

Live now, refreshed from the source and ready to deliver.

See everything live
Live

Economy and growth · World Bank

GDP

Gross domestic product of every economy, in current US dollars.

118T
World · 2025
current US$
History
Since 1960
Series
257 series
Updated
annual
Licence
CC BY 4.0
Live

Population · World Bank

Population

Total population of every country and region, each year since 1960.

8.22B
World · 2025
people
History
Since 1960
Series
260 series
Updated
annual
Licence
CC BY 4.0
Live

Economy and growth · World Bank

Inflation (consumer prices)

Annual change in consumer prices, the headline inflation rate.

3.04
World · 2025
% per year
History
Since 1960
Series
236 series
Updated
annual
Licence
CC BY 4.0
Live

Markets and finance · Federal Reserve Bank of New York

NY Fed reference rates (SOFR, EFFR, OBFR)

SOFR, EFFR, OBFR, TGCR and BGCR with percentiles and volumes, plus SOFR averages and index, every business day.

3.90
SOFR.percentRate · Sep 28, 2026
percent, USD billions
History
Since 2018
Series
36 series
Updated
daily
Licence
Public data from the Federal Reserve Bank of New York
Live

Economy and growth · World Bank

GDP per person

GDP divided by population, the most used measure of average income.

14.4K
World · 2025
current US$
History
Since 1960
Series
257 series
Updated
annual
Licence
CC BY 4.0
Live

Jobs and labour · World Bank

Unemployment rate

Share of the labour force without work but looking for it (ILO modelled estimate).

4.79
World · 2025
% of labour force
History
Since 1991
Series
230 series
Updated
annual
Licence
CC BY 4.0

From the source to your desk.

Four steps. About a minute.

  1. 1

    Find it

    Search the world's most popular datasets, from GDP to Common Crawl, or ask for the one you need.

  2. 2

    Shape it

    Pick the countries, series and history. Columns or rows: your choice.

  3. 3

    Schedule it

    Down to the minute, in your time zone, on the right business calendar.

  4. 4

    Receive it

    Checked before every delivery, with a receipt every time. Alerts if a source is late.

Your final destination.

Wherever the work happens. Email, Teams and Slack are live today; the rest are on the way.

Messages

  • EmailLive
  • Microsoft Teams chat or channelLive
  • Slack channelLive

Spreadsheets

  • Google SheetsSoon
  • Excel on OneDrive or SharePointSoon

Microsoft 365

  • SharePoint document librarySoon
  • SharePoint listSoon
  • OneDrive folderSoon

Files and storage

  • SFTPSoon
  • Amazon S3Soon
  • Azure Blob StorageSoon
  • Google Cloud StorageSoon
  • Google Drive folderSoon
  • Box folderSoon
  • Dropbox folderSoon

Databases

  • PostgreSQLSoon
  • Microsoft SQL ServerSoon
  • MySQLSoon
  • OracleSoon

Data warehouses

  • SnowflakeSoon
  • DatabricksSoon
  • BigQuerySoon

Apps and APIs

  • WebhookSoon
  • REST APISoon
  • Power BI datasetSoon

Delivery you can stake a morning on.

Checked before it leaves

Every delivery passes freshness, completeness and format checks. Bad data never ships quietly.

A receipt, every time

What was sent, where, when and which values: kept for audit, one click away.

Told, not surprised

If a publisher is late or a delivery fails, you hear about it first, in the channel you chose.

Your time zone, your calendar

Any time zone. Holidays for the Fed, NYSE and US government, or plain weekdays.

Your first feed is on us.

Pick a dataset, pick a destination, pick a time. We'll take it from there.