Skip to content
HTMLPick

How a template gets here

Found automatically, checked, then approved by a person.

Nothing is listed because it exists. An importer finds candidates, rules reject most of them, and a human approves what is left.

Browse templates
Delivery path 3 steps
  1. 01

    Find the candidates

    Read the sources every six hours.

  2. 02

    Reject most of them

    Apply the rules that decide what is actually a template.

  3. 03

    Approve it by hand

    A person looks at the screenshot and decides.

The process

Three steps, and what each one leaves you with.

  1. Step 1

    Find the candidates

    You do
    Read the sources every six hours.
    It does
    The GitHub search API by topic pair and text, plus curated registries — Hugo's themes list, the Jamstack theme index, the awesome-lists. Topics overlap heavily, so the same template usually arrives several times.
    You get
    One entry per template, matched on its canonical repository identity rather than on its name.

    01

    Find the candidates

  2. Step 2

    Reject most of them

    You do
    Apply the rules that decide what is actually a template.
    It does
    Archived repositories are out. So is anything whose source has not changed in two years, anything below the star floor, and anything whose demo link no longer answers. A topic search returns tutorials and dead experiments alongside real work, and between a third and a half of raw hits do not survive this.
    You get
    A queue of candidates rather than a dump of search results.

    02

    Reject most of them

  3. Step 3

    Approve it by hand

    You do
    A person looks at the screenshot and decides.
    It does
    The score orders the queue; it does not make the decision. An approved listing still stays out of the grid until it has an image, because a card with a hole in it is worse than a listing nobody sees.
    You get
    A page with a screenshot, a licence, a star count and the date its source last changed.

    03

    Approve it by hand

What we do not do

We do not copy the author's words.

A description here is written from the template, not lifted from its README.

A README is fetched so a machine can read it — to work out whether this is really a template, which framework it uses and what kind of site it makes. What gets stored is that derived summary, not the author's prose. Licence facts are checked against deps.dev, which publishes them under CC-BY 4.0 with caching expressly permitted, and is credited in the footer. Stars, dates and URLs are facts and are safe to republish. The one thing we are careful about is the screenshot: it reproduces the author's design, so it is rendered only when an OSI licence is detected.

What you get on a listing page

Enough to decide without cloning it first.

  • A screenshot of the template itself, not of its repository card.
  • The licence as an SPDX identifier, or "not stated" — which is the answer.
  • The date the source last changed, and a mark when that is stale.
  • Working links to the demo and the source, with dead ones recorded rather than shown.
  • Related templates on the same framework, use case and CSS toolkit.

Next step

Submit one of your own.

A free submission gets the same dedicated page a paid one gets. We do not hold the page back as an upsell.

Submit a template Ask a question