Changes for version 1.10 - 2026-11-16

  • Images load faster, Markdown gains components, drafts and scheduled posts work, dapper check looks for accessibility and SEO problems, and dapper import brings in a WordPress, Jekyll, Hugo or Medium site. See "Upgrading from 1.0" in the README.
  • Changed: pages with draft: true, and pages dated in the future, are no longer published by dapper build or dapper deploy. dapper serve shows them, as do build, watch and check with --drafts. Set future: true in _config.yml to publish pages dated in the future.
  • Added: images in _config.yml gives every image width and height, lazy loading after the first on a page, resized copies listed in a srcset, and WebP copies in a <picture>. Resizing uses ImageMagick or libvips when one is installed, and copies are kept in .dapper/images.
  • Added: components in Markdown, without template code: callouts (note, tip, info, warning, caution, important), details, gallery, youtube, vimeo, button, figure and map. Themes and sites can add their own, or restyle the built-in ones, with templates in _layout/components.
  • Added: dapper check warns about images without alt text, skipped heading levels, a page with no language, links with no text or only "click here", and missing, duplicate or overlong titles and descriptions. check: { accessibility: false, seo: false } turns them off. Its report lists the drafts and scheduled pages it left out.
  • Added: dapper import wordpress|jekyll|hugo|medium, which brings a site's pages into the source directory, keeping their addresses, dates, categories, tags and drafts, converting what it can (captions, embeds, Liquid, shortcodes) and listing what it can't. Pages are fitted to the _schema.yml of the folder they land in: in the blog starter, posts get the blog category and their own categories become tags.
  • Changed: the default starter's layout sets <html lang="en">, and the blog starter's home page has a description.
  • Changed: an HTML page's first paragraph is its excerpt, as a Markdown page's is, so it also gives the page a description. An HTML page with template code has none.
  • Fixed: AGENTS.md and CLAUDE.md, written by dapper init, were copied into the output directory and published with the site.
  • Fixed: a page without a date (which takes its file's modified time) was given the wrong time when the site's time zone differs from the computer's, which in this release would also have held it back as scheduled.
  • Fixed: a file without front matter (in the source directory, or at the top of the project) silently replaced a page written to the same place. It is now an error, like two pages writing one file.
  • Fixed: dapper check ignored links that came before a skipped heading, treated a link to another site whose address starts with the site's own (example.com.other.org) as a page of the site, named a copied HTML file by its temporary path, and reported <img alt> as having no alt text.
  • Fixed: a code block could be closed by a shorter fence, and text that happened to match Dapper's internal placeholders was replaced.
  • Fixed: a list or map given as a category ended up in the URL, and a list given as a time zone crashed the build. A symbolic link to nothing at the top of the project stopped the build; it is now skipped with a warning. Settings set to JSON false through MCP counted as on.
  • Fixed: the MCP server's write_page wrote empty front matter that Dapper couldn't read, and followed symbolic links out of the source directory. Valid JSON that isn't a request now gets an Invalid Request error instead of a parse error.
  • Fixed: a deploy command that couldn't be started reported a nonsense exit status; a file:// template address was taken for a folder; and AGENTS.md named _data/redirects.yml even when data files are elsewhere.
  • Fixed: in dapper import, WordPress ?p= links became the home page's address and category names kept &amp;; Jekyll posts with published: false, a date with a time zone offset, no front matter, or a category with a space were read wrongly; Hugo shortcodes with escaped quotes were cut off, a ref to the home page became //, and Markdown files inside a page bundle became pages; Medium stories with no title were written as .html. The TOML reader now handles comments, \u escapes, multi-line strings, nested arrays and dotted keys.
  • Fixed: images with an upper-case extension (PHOTO.JPG) got resized copies with a lower-case extension, which srcset didn't name. Maps of places with accented or non-Latin names had a broken address, and more forms of Vimeo address are recognized.
  • The test suite grows from 220 to over 5,000 checks, covering 98% of the code.

Documentation

A publishing tool for static websites.

Modules

A publishing tool for static websites.
Explain a Dapper project to AI coding agents.
Find the problems in a site without publishing it.
Category pages, tag pages and pagination.
Callouts, videos, galleries and more in Markdown.
YAML project file defaults.
Publish a site to Amazon S3, GitHub Pages, or anywhere.
Default template filters available to all projects.
Image sizes, responsive images and lazy loading.
Bring a site in from WordPress, Jekyll, Hugo or Medium.
Starter projects for the "dapper init" command.
Let AI assistants work on a Dapper site over MCP.
Convert Markdown to HTML, with fenced code blocks.
Sitemaps, feeds, Markdown copies, llms.txt and SEO tags.
Check front matter against a _schema.yml file.
A small business starter for "dapper init --template business".
A documentation starter for "dapper init --template docs".
A portfolio starter for "dapper init --template portfolio".
Common utility functions used throughout Dapper.

Provides

in lib/App/Dapper/Utils.pm