Unverified50% confidenceOpinionExact time
HTML or rich-text formats carry a large number of redundant tags, increasing token consumption and interfering with the model's comprehension of content.
1
Sources
50%
Confidence
Long-term
Relevance
10/8/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the HTML source code of modern web pages, genuinely meaningful text content often accounts for only 10%–30% of the total character count.75% similarUnverifiedThe DOM structure of a modern webpage can contain thousands of HTML elements, and providing all elements to a large language model would overflow the context window and introduce noise.74% similarUnverifiedA typical HTML document may consume 2–5 times more tokens compared to equivalent Markdown content73% similarUnverifiedFeeding raw HTML directly to an LLM both consumes a large number of tokens and exacerbates hallucinations, making it necessary to strip boilerplate web content down to clean Markdown.73% similarUnverified从ChatGPT或Claude网页界面复制文本时，浏览器剪贴板会同时存储纯文本和富文本（HTML）两种格式，富文本版本可能包含HTML标签和CSS类名71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/988826API
curl https://kongchang.com/api/v1/knowledge/claims/988826MCP
get_claim(id=988826)