# Show HN: Jekyll-panel-TUI: TUI-style jekyll theme

> Source: <https://github.com/yedhukrishnan/jekyll-panel-tui>
> Published: 2026-11-18 16:10:13+00:00

**Key takeaway:** jekyll-panel-TUI's external data panels follow a fetch-script-to-JSON-to-include pattern where a failed fetch writes empty data and a warning rather than breaking the build, and the deploy workflow reruns the scripts on every push and once a day.

**[Live preview →](https://yedhu.me)** My blog running on this theme.

A dark, terminal-style Jekyll theme for blogs, inspired by text user interfaces and old BIOS setup screens: monospace text, a blue title bar, amber panel titles and keyboard navigation (`j`/` k` or arrows to move, Enter to open, Esc to go home).

It comes with optional live panels for GitHub contributions, WakaTime coding stats (languages, editors and operating systems) and Goodreads reading, refreshed daily by GitHub Actions, plus a writing activity panel built from your own posts.

**Note:** This theme was built with the help of [Claude](https://claude.ai).

**Home**

**About page**

1. Click **Use this template** on GitHub to create your own repo. Name it`<your-username>.github.io` , or use any name with a custom domain.
2. Edit `_config.yml` :
  - site: `title` ,`description` ,`url` ,`timezone`
  - you: `author` ,`twitter` ,`social` ,`github_username` ,`linkedin_username`
  - services: `goodreads` ,`wakatime` ,`giscus` ,`goatcounter` . Leave any you don't use empty and its panel or feature stays hidden.
3. site: 
4. Replace the sample content:
  - posts: `_posts/`
  - pages: `index.md` (home intro),`about.md`
  - data files: `_data/projects.yml` ,`_data/resume.yml` (delete it to hide the Resume tab)
  - favicon: `assets/img/favicon.svg`
5. posts: 
6. For a custom domain, add a `CNAME` file containing it.
7. In the repo's **Settings → Pages** , set**Source** to**GitHub Actions** .
8. Push to `main` . The workflow fetches the panel data, builds the site and deploys it.

Colors and fonts are CSS variables at the top of `assets/css/style.css`.

| Page | File | Notes | 
|---|---|---|
| Home | `index.md` ,`_layouts/home.html` | Post index by year, an info panel previewing the selected post, and Currently Reading | 
| Post | `_layouts/post.html` | Categories, tags, share links and giscus comments | 
| Tags | `tags.html` | Filter posts by category or tag; also linked from each post | 
| About | `about.md` ,`_layouts/page.html` | Your text, plus the live panels switched on in its front matter | 
| Resume | `resume.md` ,`_data/resume.yml` | Summary, experience, projects, education, certifications and skills | 
| Projects | `projects.html` ,`_data/projects.yml` | Not linked from the tabs; the resume links to it | 
| 404 | `404.html` |  | 

```
bundle install
bundle exec jekyll serve      # http://localhost:4000
```

The live panels (GitHub, WakaTime, Goodreads) read data files that are generated at build time and not committed. To see them locally, run the fetch scripts first:

```
ruby _scripts/fetch_goodreads.rb
ruby _scripts/fetch_wakatime.rb
GITHUB_TOKEN=$(gh auth token) ruby _scripts/fetch_github.rb
```

Without them the site still builds, and those panels are hidden.

Every panel that shows outside data follows the same pattern:

1. A Ruby script in `_scripts/` fetches the data and writes a JSON file to`_data/` .
2. An include in `_includes/` renders it with the site's own styles. If the data file is empty, the panel hides itself.
3. The deploy workflow (`.github/workflows/pages.yml` ) runs the scripts before`jekyll build` , on every push and once a day, so the numbers stay fresh.

The scripts never fail the build. If a fetch fails, they write empty data and print a warning in the GitHub Actions run, so a broken source is easy to spot.

The generated files are listed in `.gitignore`, so they're never committed.

| Panel | Where | Script → data file | Include | Needs | 
|---|---|---|---|---|
| Currently reading | Home | `fetch_goodreads.rb` →`currently_reading.json` | `_layouts/home.html` | Goodreads user id | 
| Contributions | About | `fetch_github.rb` →`github.json` | `github_stats.html` | GitHub username, token | 
| Toolbox | About | `fetch_wakatime.rb` →`wakatime.json` | `wakatime.html` | WakaTime share links | 
| Writing | About | none, built from your posts | `writing_stats.html` | nothing | 
| Reading | About | `fetch_goodreads.rb` →`reading.json` | `reading_stats.html` | Goodreads user id | 

The About page panels are switched on in `about.md`'s front matter (` github_stats`, `wakatime`, `writing_stats`, `reading_stats`). Remove a line to hide that panel.

`_scripts/fetch_github.rb` uses the GitHub GraphQL API to get your contribution calendar, streaks, repo and star counts, followers and top languages. Set `github_username` in `_config.yml`.

It needs a token. In GitHub Actions the built-in `GITHUB_TOKEN` is enough, so there's nothing to set up. Locally, use `gh auth token`.

`_scripts/fetch_wakatime.rb` reads WakaTime's public share links, so it needs no API key.

1. In WakaTime, go to **Share → Embeddable Charts** (wakatime.com/share/embed).
2. Choose **Format: JSON** and a chart (Languages, Editors, ...), and copy the URL.
3. Add it under `wakatime:` in`_config.yml` , keyed by the list name:

```
wakatime:
  languages: https://wakatime.com/share/@<you>/<id>.json
  editors: https://wakatime.com/share/@<you>/<id>.json
  operating_systems: https://wakatime.com/share/@<you>/<id>.json
  hide:          # dropped from every list
    - Text
  allow:         # optional: the only entries a list may show
    editors:
      - Claude Code
      - Neovim
  rename:        # optional: show a WakaTime name under a different label
    Mac: macOS
```

WakaTime's own "Other" entry (time it couldn't classify) and anything under `hide` are dropped, and the remaining entries are rescaled to 100%. With `allow`, anything not named (browsers, Slack, ...) is dropped as well.

The panel shows the `languages`, `editors` and `operating_systems` lists. Each list shows the top 8 entries and folds the rest into "Other".

Privacy note: these lists update automatically, so any new language or editor you use will appear. Use `hide` for anything you'd rather keep private, or exclude private projects in WakaTime itself. The links only expose the chart they belong to, not your dashboard.

`_includes/writing_stats.html` is pure Liquid: it counts your posts by month and shows a year-by-month grid (only years with posts), plus totals like words written and your busiest year. There's nothing to configure.

Set `goodreads.user_id` in `_config.yml`. It's the number in your profile URL: `goodreads.com/user/show/<user_id>-name`.

Goodreads retired its shelf RSS feeds in October 2026 and has no public API, so `_scripts/fetch_goodreads.rb` reads two public pages that need no sign-in:

- **Currently reading:** the bookshelf widget, the embed code Goodreads offers for blogs.
- **Reading:** your "Year in Books" pages, one per year (`goodreads.com/user/year_in_books/<year>/<user_id>` ), for books and pages per year and recent finishes.

These are regular web pages, not official feeds, so a Goodreads redesign can break them. If that happens, the panels hide and the build shows a "Goodreads" warning.

Only books with a "date read" are counted in Year in Books.

- **Comments:**[giscus](https://giscus.app) , backed by GitHub Discussions. Set up giscus for your repo and copy its values into`giscus:` in`_config.yml` . Comments are on for posts by default; set`comments: false` in a post's front matter to turn them off.
- **Analytics:**[GoatCounter](https://www.goatcounter.com) . Set`goatcounter:` to your site code, or leave it empty to turn it off. It only loads in production builds.
- **Share links and RSS:** share links are added to each post (`_includes/share.html` ). The feed is at`/feed.xml` (from`jekyll-feed` ) and linked in the footer.
- **SEO and sitemap:**`jekyll-seo-tag` and`jekyll-sitemap` .
