You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

6.8 KiB

The Gemtexter blog engine and static site generator

This is the source code of my personal internet site and blog engine. All content is written in Gemini Gemtext format, but the script gemtexter generates multiple other static output formats from it. You can reach the site(s)...

Have a look at the content-* branches of the Git project for static content examples.

External Licenses

Gemtexter uses some external TrueType fonts for the HTML output. For license information please look into all font sub-directories of the HTML extras folder. But to summarize, all fonts are free for personal use.

Getting started


These are the requirements of the gemtexter static site generator script:

  • GNU Bash 5.x or higher
  • ShellCheck installed
  • GNU Sed
  • GNU Date
  • GNU Grep
  • Git

The script is tested on a recent Fedora Linux. For *BSD or macOS, you would need to install GNU Sed, GNU Date, GNU Grep and a newer version of Bash.


So you want such a pretty internet site too?

To get started, clone this repo and run ./gemtexter. You will be prompted with further instructions.

You will notice soon that all site content is located in ../ (you can configure the $BASE_CONTENT_DIR in gemtexter.conf). There is one sub-directory per output format, e.g.:


Alternative config file path

If you don't want to mess with gemtexter.conf, you can use an alternative config file path in ~/.config/gemtexter.conf, which takes precedence if it exists. Another way is to set the CONFIG_FILE_PATH environment variable, e.g.:

export CONFIG_FILE_PATH=~/.config/my-site.geek.conf
./gemtexter --generate

What is what

Whereas you only want to edit the content in the gemtext folder directly. The gemtexter then will take the Gemtext and update all other formats accordingly. Summary of what is what:

  • gemtext: The Gemini Gemtext markup files of the internet site.
  • html: The XHTML version of it.
  • md: The Markdown version of it.
  • meta: Some metadata of all Gemtext blog posts. It's used by gemtexter internally for Atom feed generation.
  • cache: Some volatile cache data for speeding up Atom feed generation.

Special HTML configuration

You will find the ./extras/html/header.html.part and ./extras/html/footer.html.part files, they are minimal template files for the HTML generation. There's also the ./extras/html/style.css for HTML.

Special HTML configuration

gemtexter will never touch the ../$BASE_CONTENT_DIR/html/.domains, as this is a required file for a Codeberg page. Furthermore, the robots.txt file won't be overridden as well.

Special Markdown configuration for GitHub pages

gemtexter will never touch the ../$BASE_CONTENT_DIR/md/_config.yml file (if it exists). That's a particular configuration file for GitHub Pages. gemtexter also will never modify the file ../$BASE_CONTENT_DIR/md/CNAME, as this is also a file required by GitHub pages for using custom domains.

Store all formats in Git

I personally have for each directory in ../ a separate Git repository configured. So whenever something changes I commit and push the content to Git. Gemtexter automatically detects whether a content directory is in Git or not (e.g. directories ../*/.git exist). In this case you can use the ./gemtexter --git-add command to add all files to Git and ./gemtexter --git-sync to sync all content files with the remote repository (which is a Git pull followed by a push). A ./gemtexter --git will do both, adding and syncing. Hou can set the GIT_COMMIT_MESSAGE environment variable for a custom commit message, e.g.: GIT_COMMIT_MESSAGE='New blog post' ./gemtexter --git. There's really no need to keep the cache directory in Git.

Publishing a blog post

All that needs to be done is to create a new file in ./gemtext/gemfeed/YYYY-MM-DD-article-title-dash-separated.gmi, whereas YYYY-MM-DD defines the publishing date of the blog post.

A subsequent ./gemtexter --generate will then detect the new post and link it from $BASE_CONTENT_DIR/gemtext/gemfeed/index.gmi, link it from the main index $BASE_CONTENT_DIR/gemtext/index.gmi, and also add it to the Atom feed at $BASE_CONTENT_DIR/gemtext/gemfeed/atom.xml. The first level 1 Gemtext title (e.g. # Title) will be the displayed link name. YYYY-MM-DD will be the publishing date. Various other settings, such as Author, come from the gemtexter.conf configuration file.

Once all of that is done, the gemtexter script will convert the new post (plus all the indices and the Atom feed) to the other formats, too (e.g. HTML, Markdown).

You can also have a look at $BASE_CONTENT_DIR/meta/gemfeed. There is a metafile for each blog post stored. These metafiles are required for the generation of the Atom feed. You can edit these metafiles manually and run ./gemtexter --generate or ./gemtexter --feed again if you want to change some of the Atom feed content.

Ready to be published

After running ./gemtexter --generate, you will have all static files ready to be published. But before you do that, you could preview the content with firefox ../ or glow ../ (you get the idea).

Have also a look at the generated atom.xml files. They make sense (at least) for Gemtext and HTML.

If you use git, you can use ./gemtexter --publish, which does a --generate followed by a --git-add and a --git-sync.

It is up to you to set up a Gemini server for the Gemtext, a Webserver for the HTML or a GitHub page for the Markdown format (or both).

Content filter

Once your capsule reaches a certain size it can become annoying to re-generate everything if you only want to preview one single content file. The following will add a filter to only generate the files matching a regular expression:

./gemtexter --generate '.*hello.*'

This will help you to quickly review the results once in a while. Once you are happy you should always re-generate the whole capsule before publishing it!