DEV Community: Asilbek Tilavov

From 40/100 to 100/100 AI-readability: two real before/after cases (llms.txt + JSON-LD)

Asilbek Tilavov — Thu, 03 Sep 2026 06:50:30 +0000

Two weeks ago I published a checklist for making sites readable to AI crawlers. A few people asked the fair question: does any of it actually change what AI assistants say? So here are two before/after cases from my own projects, with live files you can inspect.

Case 1: comiclens.cloud — 40/100 → 100/100

A comic-reading app. Before: no llms.txt, no structured data, robots.txt silent on AI bots. When you asked assistants about it, they either knew nothing or guessed wrong.

What shipped (about an hour, mostly copywriting):

robots.txt: explicit Allow for GPTBot, ClaudeBot, PerplexityBot
llms.txt: one markdown page of hard facts — what the app does, platforms, pricing model, feature list (live file)
JSON-LD: SoftwareApplication schema with real attributes

After re-crawl: engines that previously answered "I don't have information about this" started describing the product accurately and quoting the fact sheet.

Case 2: filmaro.app — 85/100 → 100/100

Already decent (good semantic HTML got it to 85). The gap was verifiability: no single canonical fact source, and the FAQ wasn't marked up.

Added llms.txt (live) and FAQPage JSON-LD
Rewrote vague marketing lines into checkable statements (numbers, platforms, dates)

The FAQ markup mattered most — assistants answer in Q&A format, and FAQPage schema hands them ready-made pairs.

What I'd generalize from this

The floor is lower than you think. Most sites fail at step zero: the crawler is blocked or the site renders everything client-side with no fallback text.
Facts beat prose. Every line an assistant actually quoted was a concrete fact: a year, a price, a platform name. Zero slogans survived.
Re-check after model updates. Answers drifted between engine versions during the two weeks — what Perplexity says is not what Gemini says.

Verify it yourself

Both scores above come from the scoring tool I built, AnswerRank — it asks ChatGPT and Perplexity your buyers' questions and shows whether your site is in the answers (free, no signup). But you don't need any tool to sanity-check: open ChatGPT with browsing, ask "What do you know about {your product} at {your URL}?", ship the checklist, ask again in a week. The delta speaks for itself.

Make your site readable to AI crawlers: robots.txt, llms.txt and JSON-LD in 20 minutes

Asilbek Tilavov — Wed, 26 Aug 2026 12:03:29 +0000

Your site probably has great SEO for Google — and is completely opaque to the AI assistants your customers now ask first. The fix takes about 20 minutes. Here's the whole checklist with copy-paste snippets.

1. Unblock AI crawlers (2 min)

Check yoursite.com/robots.txt. Many templates silently block AI bots. Make sure these are allowed:

User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: PerplexityBot
Allow: /

Sitemap: https://yoursite.com/sitemap.xml

If an AI crawler can't read you, nothing else matters.

2. Publish llms.txt (10 min)

llms.txt is a markdown file at your web root — a curated summary AI systems can quote. Think robots.txt, but for meaning instead of permissions:

# Acme Dental — family dentistry in Austin, TX

> Family dental clinic in Austin since 2011. Same-day emergency
> appointments, transparent pricing from $90.

## Facts
- Founded: 2011
- Services: cleanings, implants, emergency care
- Address: 12 Main St, Austin, TX
- Price range: $90–$1,200

## Pages
- [Services](https://acmedental.com/services): full price list
- [Book](https://acmedental.com/book): online booking

Rules: concrete facts only. AI engines quote specifics, never slogans.

3. Add JSON-LD (5 min)

Machine-readable truth in your <head>:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Acme Dental",
  "address": { "@type": "PostalAddress", "addressLocality": "Austin", "addressRegion": "TX" },
  "priceRange": "$90-$1200",
  "foundingDate": "2011"
}
</script>

For SaaS use SoftwareApplication, for content use Article, and add FAQPage schema to your FAQ — assistants live in question-answer format.

4. Verify (3 min)

curl yoursite.com/llms.txt → 200
Google Rich Results Test → JSON-LD parses
Ask ChatGPT: "What do you know about {your business} at {your URL}?" — with browsing on, it should now cite your facts.

The measuring problem

The annoying part is step 5: knowing whether any of this changed what AI engines actually say about you — answers vary by phrasing, region and model version. Full disclosure: that's the problem I built AnswerRank around — it asks ChatGPT/Claude/Perplexity/Gemini your buyers' questions and tracks whether you're in the answers (free check, no signup). But everything above works with or without it. Ship the files first.

Turn a Boring Book Chapter or Long Story Into a Comic You'll Actually Read (and Sell)

Asilbek Tilavov — Fri, 07 Aug 2026 06:13:23 +0000

We've all abandoned a book. Not because the story was bad — because the text was long, dense, and the evening was short. Chapters blur, classics gather dust, that 40-page story a friend sent still isn't read.

What if the text itself became a comic? Same story, same words carrying it — but as panels you actually get through.

Here's the honest, no-drawing-skills version of how to turn long text into a comic — and where to sell the result.

Disclosure: I build ComicLens. It takes the text of book pages and long stories and turns it into a comic. I'll reference it where relevant, but the workflow below is tool-agnostic.

Step 1 — Start with the text, not a blank canvas

This is the whole point: the input is writing. A chapter, a long short-story, dense pages, a wall of narration you wish were lighter.

Pick a self-contained chunk first — one chapter or one story, not a whole novel. A tight unit is easier to turn into a satisfying sequence, and easier to judge whether the result works.

Paste the text in. In ComicLens that's the core flow: long text goes in, a comic comes out. No drawing, no storyboarding by hand.

Step 2 — Let the text become a sequence of panels

A comic isn't an illustrated paragraph — it's a sequence. Good text-to-comic doesn't screenshot sentences; it finds the beats: the moment something changes, the line of dialogue that matters, the image the prose is really pointing at.

The detail that separates a real comic from a slideshow: consistency. The same character has to look like the same character three panels later; the same room has to stay the same room. If faces and places drift every panel, dense text just becomes confusing pictures. Whatever tool you use, check consistency across a full chunk before you trust it with a whole chapter.

Step 3 — Choose one style and commit

Don't mix styles inside one work. Pick one and let the whole thing live in it:

Manga / anime — huge audience, great for serialized long stories
Classic ink & halftone — timeless, reads as "real comic"
Storybook — soft and warm, perfect for lighter or younger material
Noir / cinematic — strong for tense, moody chapters

One consistent style is what turns "AI output" into something that looks published.

Step 4 — Make the words earn their place

You're adapting text, so the temptation is to keep all of it. Don't.

Let the art carry the description. If a panel shows a storm, you don't need the sentence describing the storm.
One idea per balloon. Keep dialogue; trim narration the picture already tells.

Adapting is cutting. The comic should read faster than the prose, or you haven't adapted — you've just captioned.

Step 5 — Now sell it (legally)

Turning text into comics opens a real product — but mind whose text it is:

Sell only what's yours to sell: your own writing, or public-domain works (older classics whose copyright has expired). Don't sell a comic of someone else's in-copyright book.

With that settled, here's where comics actually sell in 2026:

Direct-to-fan (highest margin)

Gumroad — sell a PDF or CBZ in five minutes. Great for a finished chapter or one-shot.
Ko-fi / Patreon — serialize a long story chapter by chapter; sell access to the next installment.

Print without inventory

Print-on-demand (Redbubble, Etsy + a POD partner) — turn a strong page into a poster or the whole thing into a printed book.

Marketplaces with built-in traffic

Webtoon / Tapas — serialize for free, monetize through their coin/ad systems. The discovery is done for you.

The channel most people underuse

Your best salesperson is a single page. Post one strong panel-page — "I turned a whole chapter into this" — to TikTok, Reels, or Pinterest. People don't want your comic; they want to do it to their own book or story. Watermark it, add a link, let it travel.

A realistic first-week plan

Day	Do
1	Pick one chapter or long story (yours or public-domain). Turn it into a comic.
2	Fix consistency and lock one style.
3	Export as PDF/PNG. Set up a Gumroad or Ko-fi page.
4–5	Post one page with the "whole chapter → comic" hook.
6–7	Reply to every comment. Adapt the next chapter.

The takeaway

The bottleneck was never the story — it's that long text is a slog to get through. Turning it into a comic doesn't dumb it down; it makes people actually finish it. You don't need to draw. You need one chunk of text, one style, and the discipline to adapt (cut) rather than caption.

If you want the fastest path from a wall of text to a finished comic, that's exactly what I built ComicLens for. But whatever you use: feed it real text, adapt honestly, and put the result where people can buy it.

What's the one book or long story you'd turn into a comic first? Drop it in the comments.
#ai #writing #sideproject #tutorial