Skip to main content
The ZonaTMO handler is one of the most sophisticated extractors in the project, supporting both single chapters and full series downloads using AI-powered image detection.

​
Supported URLs

The handler recognizes two URL patterns:

​
Single chapter

https://zonatmo.com/view_uploads/[chapter-id]
https://zonatmo.com/viewer/[chapter-id]
https://zonatmo.com/viewer/[chapter-id]/paginated

​
Full series (cover page)

https://zonatmo.com/library/manga/[series-name]
The handler automatically detects which type of URL you provide and adjusts its extraction strategy accordingly.

​
Extraction technology

​
Crawl4AI with Gemini LLM

ZonaTMO uses Crawl4AI with Google Gemini 1.5 Flash for intelligent image extraction: 
llm_config = LLMConfig(
    provider="gemini/gemini-1.5-flash", 
    api_token=config.GOOGLE_API_KEY
)

instruction = """Extract all image URLs. Look for 'data-original' 
and 'src'. Return JSON {'images': ['url1'...]}."""

llm_strategy = LLMExtractionStrategy(
    llm_config=llm_config, 
    instruction=instruction
)
Location: ~/workspace/source/core/sites/zonatmo.py:153-155

​
Cascade view optimization

The handler converts paginated URLs to cascade view for more efficient extraction: 
if "/paginated" in final_url:
    target_url = final_url.replace("/paginated", "/cascade")
elif "/viewer/" in final_url:
    if not final_url.endswith("/cascade"):
        target_url = final_url + "/cascade"
Location: ~/workspace/source/core/sites/zonatmo.py:139-142 Cascade view loads all images on a single scrollable page, making extraction faster and more reliable.

​
Lazy loading handling

ZonaTMO pages use lazy loading for images. The handler executes JavaScript to trigger loading: 
(async () => {
    const sleep = (ms) => new Promise(r => setTimeout(r, ms));
    window.scrollTo(0, 0);
    let totalHeight = 0; 
    let distance = 1000;
    while(totalHeight < document.body.scrollHeight) { 
        window.scrollBy(0, distance); 
        totalHeight += distance; 
        await sleep(200); 
    }
    await sleep(1000);
})();
Location: ~/workspace/source/core/sites/zonatmo.py:157-165 This script:
  1. Scrolls to the top of the page
  2. Gradually scrolls down in 1000px increments
  3. Waits 200ms between scrolls to trigger lazy loading
  4. Waits 1 second after reaching the bottom

​
Full series extraction

When you provide a cover/library URL, the handler:
  1. Crawls the series page to find all chapter links
  2. Extracts the manga title from <h1> tags
  3. Creates a folder named after the series
  4. Downloads each chapter as a separate PDF
  5. Reverses chapter order (oldest to newest)
links = re.findall(
    r'href=["\']https://zonatmo.com/view_uploads/[^"\']+)["\']', 
    result.html
)

# Remove duplicates while preserving order
clean_links = []
seen = set()
for l in links:
    if l not in seen:
        clean_links.append(l)
        seen.add(l)

clean_links.reverse()  # Oldest first
Location: ~/workspace/source/core/sites/zonatmo.py:49-56,90

​
Title extraction

The handler tries multiple methods to extract the manga title:
  1. H1 element with class element-title (preferred)
  2. Page title tag as fallback
  3. Default: "Manga_ZonaTMO" if nothing found
h1_match = re.search(
    r'<h1[^>]*class=["\'].*?element-title.*?["\'][^>]*>(.*?)</h1>', 
    result.html, 
    re.IGNORECASE | re.DOTALL
)

if h1_match: 
    raw_html = h1_match.group(1)
    # Remove <small> tags
    raw_html = re.sub(r'<small[^>]*>.*?</small>', '', raw_html, flags=re.IGNORECASE | re.DOTALL)
    manga_title = clean_filename(raw_html)
Location: ~/workspace/source/core/sites/zonatmo.py:66-74

​
Fallback extraction

If LLM extraction fails, the handler uses regex fallback: 
if not image_urls and result.html:
    matches = re.findall(
        r'(https?://(?:img1?\.?tmo\.com|otakuteca\.com|img1tmo\.com)[^"\'\s]+\.(?:webp|jpg|png))', 
        result.html
    )
    if matches: 
        image_urls = sorted(list(set(matches)))
Location: ~/workspace/source/core/sites/zonatmo.py:185-187

​
Image filtering

The handler filters out non-content images: 
image_urls = [
    u for u in image_urls 
    if "cover" not in u 
    and "avatar" not in u 
    and "banner" not in u
]
Location: ~/workspace/source/core/sites/zonatmo.py:189

​
Headers and configuration

ZonaTMO requires specific headers defined in config.py: 
HEADERS_ZONATMO = {
    "User-Agent": "Mozilla/5.0 ...",
    "Referer": "https://zonatmo.com/"
}
These headers are passed to download_and_make_pdf for image downloads.

​
Usage examples

​
Single chapter

from core.handler import process_url

await process_url(
    "https://zonatmo.com/view_uploads/123456",
    log_callback=print,
    check_cancel=lambda: False,
    progress_callback=lambda current, total: print(f"{current}/{total}")
)
Output: PDF/zonatmo_chapter.pdf (or auto-detected title)

​
Full series

await process_url(
    "https://zonatmo.com/library/manga/my-favorite-manga",
    log_callback=print,
    check_cancel=lambda: False,
    progress_callback=lambda current, total: print(f"Chapter {current}/{total}")
)
Output: 
PDF/
└── My Favorite Manga/
    ├── My Favorite Manga - 001.pdf
    ├── My Favorite Manga - 002.pdf
    └── My Favorite Manga - 003.pdf

​
Web interface

You can also use the web version:
  1. Start the server: START_WEB_VERSION.bat
  2. Navigate to http://localhost:3000
  3. Paste any ZonaTMO URL
  4. Watch real-time logs and progress bars

​
Rate limiting

The handler includes a 1-second delay between chapters to avoid rate limiting: 
for i, chap_url in enumerate(clean_links):
    await self._process_chapter(chap_url, ...)
    await asyncio.sleep(1)  # Polite delay
Location: ~/workspace/source/core/sites/zonatmo.py:105-106

​
Known limitations

Requires Google API key: The LLM extraction will fail without GOOGLE_API_KEY in your .env file. Regex fallback may not catch all images.

​
URL resolution

The handler attempts to resolve redirects before extraction: 
async with aiohttp.ClientSession() as session:
    async with session.get(url, headers=config.HEADERS_ZONATMO) as resp:
        if resp.status == 200:
            final_url = str(resp.url)
Location: ~/workspace/source/core/sites/zonatmo.py:134-137 If this fails, it uses the original URL with a warning.

​
Duplicate detection

Chapter lists may contain duplicate URLs. The handler deduplicates using a set: 
clean_links = []
seen = set()
for l in links:
    if l not in seen:
        clean_links.append(l)
        seen.add(l)
Location: ~/workspace/source/core/sites/zonatmo.py:51-56

​
Implementation details

​
Class structure

class ZonaTMOHandler(BaseSiteHandler):
    @staticmethod
    def get_supported_domains() -> list:
        return ["zonatmo.com"]
    
    async def process(self, url, log_callback, check_cancel, progress_callback):
        # Main processing logic
        ...
    
    async def _process_chapter(self, url, output_name, ...):
        # Chapter-specific extraction
        ...
Location: ~/workspace/source/core/sites/zonatmo.py:20-214

​
Domain matching

The handler is automatically selected by the routing logic in handler.py when the URL contains "zonatmo.com".

​
Next steps

TMO-H

Similar AI-powered extraction for TMO-H

Configuration

Set up your Google API key

Architecture

Learn about the Strategy Pattern

Utils

Explore download_and_make_pdf function

Build docs developers (and LLMs) love

Get started for freeTalk to us