Full audit.
56 findings.

Claude's structured read: "nikki.lol is a thoughtfully hand-crafted personal digital notebook by someone who writes to think, not to perform. It's a rare, tracker-free corner of the web filled with short reflections on AI, philosophy, learning, and daily life." Citation-worthy. Sentiment: positive. Credibility: high. Top three green flags: zero trackers, named author, plain-text honesty.

AI assistants and agents look for /llms.txt as a curated reading list of your site. Without it they crawl HTML inefficiently and may miss your most important pages. Spec at https://llmstxt.org.

# nikki.lol

> A personal notebook. Short reflections on AI, philosophy, learning.

## Notes
- [Universal mentor-coach prompt](https://nikki.lol/notes/universal-mentor-coach-prompt/): a system prompt for AI mentors
- [Dead branch trail](https://nikki.lol/notes/2026-03-24-dead-branch-trail/): on letting go of unfinished work

Only ~4% of sites declare Content-Signal in robots.txt. Without it, AI agents have no explicit signal about whether they may use your content for training, inference, or search indexing.

User-agent: *
Allow: /

Content-Signal: search=yes, ai-train=no, ai-input=yes
Sitemap: https://nikki.lol/sitemap.xml

The first content under your H1 is neither a TL;DR / summary block nor a concise lede paragraph. Pages with explicit upfront answers are quoted by ChatGPT, Perplexity, Claude, and Google AI Overviews far more often than wall-of-text intros.

Two of your primary buttons render as Unicode glyphs (↻ and ⌬) with no visible label and no `aria-label`. Visitors without context guess; screen readers announce nothing.

<button aria-label="Refresh notes" title="Refresh notes">
  ↻ <span class="visually-hidden">Refresh</span>
</button>
<button aria-label="Random note" title="Random note">
  ⌬ <span class="visually-hidden">Random</span>
</button>

The H1 just says "nikki.lol" — the domain. A first-time visitor lands and reads the URL twice. Move the value prop ("a personal notebook on AI, philosophy, and learning") into the H1 itself, or into the first sentence below it.

Best guess at your pitch from the hero: 'A personal site or notebook of some kind.' (confidence 35/100). Hero headline is just the domain name, CTAs unlabeled, no tagline beneath the title.

First-time visitors who like your writing have no way to come back. Add at least an RSS link or a one-field email subscribe — even a static "follow at @nikki" link beats nothing.

/notes/universal-mentor-coach-prompt/, /notes/2026-03-24-dead-branch-trail/, /notes/the-rise-and-fall-of-a-midwest-princess/, and /notes/2026-03-18-change-requires-consistency/ all lack a ``. Google auto-generates snippets from body text — the auto-pick is rarely your strongest sentence.

<meta name="description" content="A short reflection on letting go of unfinished work — and why some branches deserve to die.">

Search engines and AI crawlers expect a robots.txt at the site root. Add one to control crawling explicitly — even a minimal one is better than 404.

User-agent: *
Allow: /

Sitemap: https://nikki.lol/sitemap.xml

No JSON-LD or microdata detected. Without schema markup, you miss out on rich results (FAQ, HowTo, Article, Person, Breadcrumb) that visibly outperform plain blue links. For a personal-notebook site, Person and Article schema are the priorities.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Person",
  "name": "Nikki Nikkhoui",
  "url": "https://nikki.lol",
  "sameAs": ["https://github.com/nikkinikki"]
}
</script>

Sitemaps help search engines find every URL on your site quickly, especially for new pages and deep navigation. We probed /sitemap.xml, /sitemap_index.xml, /sitemap/sitemap.xml, /sitemaps/sitemap.xml — none resolved.