For developers

A free data feed for every chart

Every chart on Chartly comes with its numbers, its public source, its license and the date of its newest data. The feed gives you all of that as JSON, for your app, newsroom tool, class project or AI assistant. It's free, and you don't need a key or an account. Charts that update themselves update here too.

Credit Chartly, with a link to each chart

The feed is free on one condition: wherever you show or quote a chart's numbers, name Chartly with a link to that chart's page (its url), and credit its source as its license asks. Every chart has a ready-made credit line, and credit_html with the link in place, for example:

U.S. Census Bureau, Annual Survey of School System Finances, via Chartly (https://chart.ly/charts/school-spending-per-student-largest-districts)

Add Chartly to your AI assistant

Chartly has a free connector for AI assistants that support the Model Context Protocol (MCP), such as Claude and ChatGPT. Once it's added, your assistant can search Chartly's charts and quote a chart's latest number with its source, its date and a link to the chart, so you can check it. It can also quote the guide, statistics page or city, county, country, college or airport page that best answers your question, such as "Is crime going up in Chicago?", "What does Harvard University really cost?" or "How often are flights at Denver late?". No key or account is needed. Add it as a custom connector with this address:

https://chart.ly/mcp
  • Claude: open Settings, then Connectors, choose "Add custom connector", name it Chartly and paste the address.
  • ChatGPT: in Settings, under Apps & Connectors, turn on developer mode in the advanced settings, then create a connector named Chartly with the address.
  • Other tools: use the address as a remote MCP server over Streamable HTTP. It has two read-only tools: search (keywords) and fetch (a chart's id or a page's, such as guides/is-crime-falling, city-crime/chicago-il or airports/denver-den).

It offers the same charts as the data feed, on the same terms: an assistant that answers with a chart's or page's numbers should say their date and link to what it used. Building your own tool? Start from llms.txt, a short plain-text guide to the feed.

Every chart

Lists every published chart, newest first: its question, takeaway, topic, labels and units, source, license, data date and links. Add ?topic=health (a topic's name in lowercase, with hyphens) for one topic.

curl https://chart.ly/api/v1/charts

One chart and its numbers

The same details plus data, the points as charted: each has a label and a value, and some have a margin (a survey's margin of error, ±) or a url (the filing a figure comes from). The slug is the end of the chart's address on Chartly.

curl https://chart.ly/api/v1/charts/school-spending-per-student-largest-districts

Open this example

What each chart has

title
The question the chart answers, in plain words.
takeaway
The answer, with the key number.
x_label, y_label
What the labels and values are, with units.
unit
"dollars", "percent" or null, read from y_label. Values are never scaled: whole dollars, 12.3 for 12.3%.
source, license
Who published the numbers and the terms they may be reused under, each with a link.
data_as_of
The date of the newest data point.
updated_at
When the chart last changed on Chartly.
url, csv_url, image_url, embed_url
The chart's page, its numbers as a spreadsheet file, its picture and its embeddable version.
credit, credit_html
The line to show with the chart's data: its source and Chartly, linked to the chart's page.
annotations
Events marked on a chart over time: from, to, title and text.
average_label
The benchmark point in a ranking ("U.S. overall"), if it has one.

Using the data

  • Credit each chart's source and Chartly, with a link to the chart's page (url). This is required.
  • Follow the license each chart names. Most U.S. government data is public domain; World Bank and some other data is CC BY 4.0, which asks you to name the source.
  • Each address can make up to 60 requests a minute. Please save what you fetch and check updated_at rather than asking again and again.
  • Websites can call the feed straight from the browser.
  • A chart's address stays the same. If a chart is ever renamed, its old address redirects to the new one, so links and saved slugs keep working.
  • A lookup chart (baby names, companies, banks, ZIP codes, countries) has the numbers for the item it opens with; its page looks up the rest.

Left out of the feed

Some sources let us show their numbers on our pages but don't allow passing them on in bulk, so their charts aren't in the feed: CollegeFootballData.com, nflverse (NFL data), and Hacker News. That includes the college football and NFL charts and the Hacker News dashboard. You can still read them on Chartly.

Missing a chart?

If you need a chart we don't have yet, tell us what you're building.

Tell us what you need