Skip to content

Open source · built for public-sector teams

Pub Health Catalog Template

One repository, no server, four ways to run it.

A catalog site that lives on GitHub Pages and is managed entirely through GitHub: someone fills in a form, a maintainer merges a pull request, the site rebuilds. The four examples below are this same repository with different settings — open any of them and click around.

Live examples

Four complete sites, built from one repository

Each one is a full build with its own fields, filters, colours and sample content — search, feeds and the submission form included. Nothing here is a mockup.

  • The AI use case catalog example: a deep navy hero headed "What public-sector teams are building with AI", with a search box, a count of ten use cases, and the three newest ones listed in a panel beside it.

    AI use case catalog

    A shared catalog of AI projects

    The configuration this template ships with. Public-sector teams publish what they have built, with enough detail — the data it touches, the skills it took, the licence, the running cost — for another team to judge whether it could reuse it. Filter the catalog, shortlist three entries and compare them side by side.

    Entries are called
    Use cases
    Content model
    44 fields · 21 filters
    Turned on
    • Catalog
    • Submissions
    • Featured carousel
    • Stats
    • Governance
    Open the AI use case catalog example
  • The cohort portal example: an indigo hero headed "What our teams built with data", with a search box for team projects, the three newest listed beside it, and tiles for browsing by track and skills practiced below.

    Program / cohort portal

    A training program's portal, cohort by cohort

    Teams publish the projects they built during a fellowship or academy, tagged by cohort year, area of work and track. The cohorts and events modules add program-year pages with timelines, session materials and a shared calendar around the gallery.

    Entries are called
    Team projects
    Content model
    20 fields · 6 filters
    Turned on
    • Catalog
    • Submissions
    • Featured carousel
    • Stats
    • Events
    • Cohorts
    Open the Program / cohort portal example
  • The resource library example: a dark green hero headed "Find the resource you need, fast", with a search box, the three newest resources beside it, and tiles for browsing by resource type, audience, access and format below.

    Resource library

    A shelf of guides, toolkits and datasets

    Shorter entries, more of them. Each resource is a link, a format, an audience and an honest note on what it takes to get access — the right shape for a curated library of things other people made and you want your staff to find.

    Entries are called
    Resources
    Content model
    15 fields · 7 filters
    Turned on
    • Catalog
    • Submissions
    • Featured carousel
    • Stats
    Open the Resource library example
  • The blank catalog example: a slate hero headed "A catalog of things worth sharing", with a search box, the three newest entries beside it, and one category filter in the browse-by band below.

    Blank catalog

    The smallest useful starting point

    Title, summary, one category, tags, links and images — nothing you have to explain away to your first contributor. Rename the fields and this becomes a catalog of grants, datasets, policies, vendors or anything else worth listing in one place.

    Entries are called
    Entries
    Content model
    8 fields · 2 filters
    Turned on
    • Catalog
    • Submissions
    Open the Blank catalog example

What it does

What you get, whichever example you start from

  • The fields are yours

    One file defines every question an entry answers, and how each answer is shown — as a card badge, a filter, a sidebar fact. The forms, the issue template, the search index and the validator all follow it.

    The content model
  • GitHub is the CMS

    There is no admin login and no database. Contributors submit through a web form, maintainers review in a pull request, and the page history is the audit trail.

    The maintainer guide
  • Nothing publishes without review

    Every submission arrives as a draft with your review checklist attached. Answers you flag as sensitive — protected health information, a public-facing audience — raise the review to a second pair of eyes automatically.

  • Governance on the page, not in a PDF

    How review works, who does what and the standing policies are published from one data file. A monthly job counts submissions, publications and review turnaround from the repository itself, so the numbers are not a claim.

  • Built for deciding, not just browsing

    Cards are laid out to answer "could my team reuse this?" in a couple of seconds. Shortlist up to three entries, read them field by field, and print the comparison as a clean brief with the links spelled out.

  • Search, filters and feeds

    Search reads the whole write-up and names the section it matched in. Filters sit beside the results and restore from the URL, so a filtered view pastes into an email. New entries publish to an Atom feed.

  • Modules you can turn off

    Catalog, submissions, events, cohorts, resources, governance and the home page blocks each switch on or off. Navigation and the home page adapt, and pages under a module you turned off leave the build entirely.

  • Accessibility is a build rule

    Nothing is signalled by colour alone, every control has a visible focus ring, and the catalog still works with JavaScript off. Automated WCAG 2.1 AA and performance checks run on every change.

  • Nothing to host, nothing to patch

    GitHub Pages serves plain files, so there is no server to pay for, secure or migrate. Template improvements reach your copy when you choose to merge them, and your content and settings are never overwritten.

    How upgrades work

How publishing works

From a form to a published page in three steps

The same path for every entry, on every example. No CMS login, no database, no one editing HTML.

  1. Someone fills in a form

    A contributor answers the questions your schema defines, on a page that previews the card their entry will produce. Screenshots are dragged straight onto it. All they need is a GitHub account.

  2. Automation drafts the pull request

    A workflow reads the submission, saves the images into the entry's folder and opens a pull request with the entry already written, the review checklist filled in, and a flag on anything that needs closer scrutiny.

  3. A maintainer merges it

    The reviewer checks the entry against your published criteria and merges. The site rebuilds and the entry is live a couple of minutes later. Corrections work the same way, from a link on the page.

Start here

Ready to try it with your own content?

  • Follow the launch guide

    The full tutorial: create the repository, turn on the three settings GitHub needs, configure the site, clear the samples and publish a first entry end to end.

  • Configure it in your browser

    The setup wizard, no terminal required. Start from one of the four presets, change the names, colours and fields, and it writes the configuration files for you.

  • Read the documentation

    Every guide and who it is for — maintainers reviewing submissions, contributors submitting them, and whoever is retargeting the template at a different subject.