diff --git a/changedetectionio/html_tools.py b/changedetectionio/html_tools.py
index 824c0e6aa..6b93402e2 100644
--- a/changedetectionio/html_tools.py
+++ b/changedetectionio/html_tools.py
@@ -565,6 +565,12 @@ def html_to_text(html_content: str, render_anchor_tag_content=False, is_rss=Fals
if is_rss:
html_content = re.sub(r'])', r'', r'', html_content)
+ else:
+ # Strip bloat in one pass, SPA's often dump 10Mb+ into the for styles, which is not needed
+ # Causing inscriptis to silently exit when more than ~10MB is found.
+ # All we are doing here is converting the HTML to text, no CSS layout etc
+ html_content = re.sub(r'<(?:style|script|svg|noscript)[^>]*>.*?|<(?:link|meta)[^>]*/?>|',
+ '', html_content, flags=re.DOTALL | re.IGNORECASE)
text_content = get_text(html_content, config=parser_config)
return text_content
diff --git a/changedetectionio/tests/unit/test_html_to_text.py b/changedetectionio/tests/unit/test_html_to_text.py
index f4787a47c..9815c27a1 100644
--- a/changedetectionio/tests/unit/test_html_to_text.py
+++ b/changedetectionio/tests/unit/test_html_to_text.py
@@ -199,6 +199,91 @@ class TestHtmlToText(unittest.TestCase):
print(f"✓ Basic thread-safety test passed: {len(results)} threads, no errors")
+ def test_large_html_with_bloated_head(self):
+ """
+ Test that html_to_text can handle large HTML documents with massive bloat.
+
+ SPAs often dump 10MB+ of styles, scripts, and other bloat into the section.
+ This can cause inscriptis to silently exit when processing very large documents.
+ The fix strips \n'
+
+ # Generate massive script block (~5MB)
+ large_script = '\n'
+
+ # Generate lots of SVG bloat (~3MB)
+ svg_bloat = '\n' * 50000
+
+ # Generate meta/link tags (~2MB)
+ meta_bloat = '\n' * 50000
+ link_bloat = '\n' * 50000
+
+ # Generate HTML comments (~1MB)
+ comment_bloat = '\n' * 50000
+
+ # Generate noscript bloat
+ noscript_bloat = '\n' * 10000
+
+ # Build the large HTML document
+ html = f'''
+
+
+ Test Page
+ {large_style}
+ {large_script}
+ {svg_bloat}
+ {meta_bloat}
+ {link_bloat}
+ {comment_bloat}
+ {noscript_bloat}
+
+
+
Important Heading
+
This is the actual content that should be extracted.
+
+
First paragraph with meaningful text.
+
Second paragraph with more content.
+
+
+
+
+'''
+
+ # Verify the HTML is actually large (should be ~20MB+)
+ html_size_mb = len(html) / (1024 * 1024)
+ assert html_size_mb > 15, f"HTML should be >15MB, got {html_size_mb:.2f}MB"
+
+ print(f" Testing {html_size_mb:.2f}MB HTML document with bloated head...")
+
+ # This should not crash or silently exit
+ text = html_to_text(html)
+
+ # Verify we got actual text output (not empty/None)
+ assert text is not None, "html_to_text returned None"
+ assert len(text) > 0, "html_to_text returned empty string"
+
+ # Verify the actual body content was extracted
+ assert 'Important Heading' in text, "Failed to extract heading"
+ assert 'actual content that should be extracted' in text, "Failed to extract paragraph"
+ assert 'First paragraph with meaningful text' in text, "Failed to extract first paragraph"
+ assert 'Second paragraph with more content' in text, "Failed to extract second paragraph"
+ assert 'Footer text' in text, "Failed to extract footer"
+
+ # Verify bloat was stripped (output should be tiny compared to input)
+ text_size_kb = len(text) / 1024
+ assert text_size_kb < 1, f"Output too large ({text_size_kb:.2f}KB), bloat not stripped"
+
+ # Verify no CSS, script content, or SVG leaked through
+ assert 'color:red' not in text, "Style content leaked into text output"
+ assert 'console.log' not in text, "Script content leaked into text output"
+ assert ' {text_size_kb:.2f}KB text")
+
if __name__ == '__main__':
# Can run this file directly for quick testing