npm downloads
UnavailableRolling 30 daysNot currently measuredSourceThe EmDash Observatory
EmDash adoption statistics:
Is the growth sticking?
EmDash launched in April 2026. Downloads tell us people are paying attention. They don't tell us how many sites are running it. Here's what the evidence shows, what it doesn't, and how the numbers are changing.
GitHub stars
UnavailableObserved snapshotNot currently measuredSourceCore CMS releases
UnavailableVerified GitHub releasesNot currently measuredSourceDays of npm history
UnavailableRecorded daily observationsNot currently measuredSourceRefresh schedule: npm and GitHub daily, WordPress monthly. Last verified observation dates appear above. Automatic collection begins when the Cloudflare Worker is deployed.
EmDash growth dashboard
Are EmDash downloads holding up?
Daily npm data, not a growth curve drawn for a pitch deck. The orange markers show when core EmDash releases landed, so you can see what happened next.
Loading verified download history…
Data source: npm public download counts. Historical dates are UTC.
Orange markers indicate published core EmDash releases. Hover or focus each marker for its version and date.
On a narrow screen, scroll sideways to inspect the chart.
See the download figures behind this chart
| Date (UTC) | Value |
|---|---|
| No collected data loaded yet. | |
The adoption question
A launch spike is easy. Staying power is harder.
EmDash caught developers' attention when it launched. That matters. But curiosity isn't a deployment, and an npm request isn't a live site.
The better question: are downloads holding up between releases, or are we seeing short bursts of experimentation? A giant growth percentage means little without that context.
First observed daily count
UnavailableThe starting point, measured in daily npm package downloads. Not installations.
Latest observed daily count
UnavailableThis is one day's npm traffic. Use the rolling chart to judge the trend.
Growth comparison
EmDash vs Astro: who's gaining momentum?
Here's a comparison with the same unit on both sides: rolling 30-day npm downloads. Indexed growth shows which package is accelerating from a shared starting point. Net change and absolute downloads show the actual numbers on separate scales, because Astro is much larger. WordPress gets its own website-share chart below.
Loading comparable history…
Same dates. Two vertical scales. Each series has its own labelled graph. Compare the values, not the line heights.
On a narrow screen, scroll sideways to inspect each chart.
Indexed mode starts each series at 100 on the shared positive observation date. Downloads are package requests, not websites.
View comparison data tableExact dated values, source units and missing observations
These are the measurements behind the chart, matched by date. Switch the comparison mode above to see the same readings indexed, as net change, or as absolute downloads.
| Date (UTC) | EmDash index | Astro index |
|---|---|---|
| No comparable observations loaded yet. | ||
A closer look at WordPress
WordPress still dominates. Its website share is slipping.
A few percentage points can disappear on a chart alongside fast-growing npm packages. Here, WordPress has its own scale. Every point is a dated W3Techs website-share observation.
Loading WordPress measurements…
The vertical axis is zoomed to show small changes. This is website share, not npm downloads. Lines join published dates; they do not imply measurements in between.
View WordPress source observations
| Observation date | Website share |
|---|---|
| No observations loaded yet. | |
Different metrics, different questions. An npm download isn't a website, and WordPress's surveyed website share isn't a package count. Check the W3Techs source ↗
Scale versus momentum
Growing fast isn't the same as being big.
Fast growth makes a good headline. Scale tells you something else. Astro has a far larger npm footprint; WordPress is measured by its share of surveyed websites. Neither is a shortcut to counting EmDash installations.
A decline isn't a migration
Unavailable to Unavailable of surveyed websites, from first verified reading to latest verified reading.
That's a change of an unknown number of percentage points. It tells us nothing about whether those websites moved to EmDash.
W3Techs tracks share of all surveyed websites. Its CMS-only market share uses a different denominator again.
See the dated W3Techs series ↗The download gap is real
Latest complete 30-day npm counts: Unavailable for EmDash and Unavailable for Astro.
That's an Astro-to-EmDash ratio of roughly unavailable. These are packages with different roles. The ratio describes download volume, not CMS market share.
Period ending when data is available (UTC).EmDash timeline
Did releases move the needle?
Release dates are facts. Their impact takes a bit more work. Compare each core CMS release with the download trend, but don't assume the two are related just because they happened close together.
- EmDash beta announced
Cloudflare introduced EmDash as an Astro-based CMS. Announcement ↗
- EmDash 1.0 and plugin registry
Stable CMS and decentralised plugin registry announced. Announcement ↗
Read this before quoting a chart
What these numbers can't tell you.
Downloads aren't live sites
Build servers, repeat installs and experiments all create npm requests. None tells us how many websites actually run EmDash.
A star isn't a customer
Someone starred the repository. Good. They might also have bookmarked it for later. We record actual daily snapshots, not an invented history.
Website detectors miss things
Private and undiscoverable installations won't necessarily show up in a census. An estimate without a stated method and date isn't much use.
WordPress losing share proves one thing
Its measured share fell over the period shown. It doesn't prove that a single website switched to EmDash.
Read the small print. It matters.
Where the numbers come from.
If we're going to put numbers in a headline, you'd better be able to check them. Here's the source, the maths, and what's missing.
Sources, formulas and coverageMethods, definitions, source links and refresh policy
npm downloads
Daily npm download records for emdash and astro, starting in April 2026 where available. Build servers, reinstallations and experiments all count. Websites do not.
7-day average: seven consecutive daily counts divided by seven. Rolling 30 days: thirty consecutive daily counts added together. Missing days never get invented.
npm data method ↗GitHub snapshots
Stars and forks are observed from the official public repositories. A star is a signal of interest, not evidence of a customer or installation.
Our star history starts when collection starts. We don't fill earlier gaps with guesses. Core EmDash releases use their published GitHub dates.
GitHub API method ↗WordPress website share
WordPress figures show its share of all surveyed websites, not just websites with a known CMS. No downloads are involved.
Only dated, published observations appear. The dashed line joins monthly readings; it does not represent measured values between dates. Website share has its own chart because the unit differs from npm downloads.
W3Techs historical readings ↗Indexed growth
The two npm series both start at 100 on the first date with positive readings for each. That shows relative movement. It does not make their total downloads equal.
Net change uses the same baseline for EmDash and Astro npm downloads. WordPress uses a separate percentage scale and is never mixed into their Y-axis.
Updated automatically. Never guessed.
The collector is scheduled to check npm and GitHub daily, and WordPress monthly. If a source fails, we keep its last verified reading and date. Plugin totals and detected EmDash site counts stay blank until we have a source worth trusting.
The last confirmed observation date appears when the data pipeline is connected.
Want to follow the plugin ecosystem? Official plugin registry documentation ↗
FREE GUIDE + AUDIO EDITION
Considering EmDash?
Read this before you build.
The numbers tell you how much attention EmDash gets. The free Ultimate Guide explains what the CMS does, where it fits, and what to consider before using it. Read it or listen to the audio edition.
Explore the free guideFree to read or listen. No chart-reading required.
Build with EmDash
Curious enough to try EmDash?
A chart can't tell you whether a CMS fits your project. A working site can. Build one free with DashHarbor and find out.