Idiomas

Pre-publish Comparison

Free, no signup. Compare a draft HTML export with the current live HTML you supply before approving a release.

Built by Patrick Stox

Supply both HTML documents

1 MiB maximum per document. Oversized input is rejected in full; it is never silently shortened. Use HTML only—do not paste credentials, access tokens, personal data, or confidential source comments.

Runs entirely in your browser — nothing you paste is uploaded or stored. The tool does not fetch URLs, call AI, use credentials, persist input, or include source HTML in exports. Anonymous run-level outcome counters may be used for aggregate research; URLs, domains, IPs, and identifiers are never included, and no statistic is released below 100 runs.

Feedback
Report a bug

Found something broken in Prepublish Comparison? Let us know what happened — this goes straight to a private triage queue, not a public list.

What will be sent 
 No tool inputs, uploads, pasted source, complete results, query parameters, or URL fragments are attached automatically. You can edit or remove the selected passage above. Browser and anti-abuse metadata is processed for spam prevention. 
  
 
  
 
 
  
 
 
   
 
 
 
Sample report
 
Fabricated draft and live HTML
 
Regression: the draft removes one live hreflang declaration.
Changed: the title and heading outline differ.
Not evaluated: AI-crawler posture is not observable from the supplied HTML.
 
 
 
How to use it
 
  1. Export or view the complete HTML for the draft and current live page.
  2. Paste both documents or choose two local .html files. Inputs over 1 MiB are rejected.
  3. Compare the normalized signal rows. Investigate regressions first, then confirm whether each changed row is intentional.
  4. Download CSV or JSON rows for a release checklist. The exports omit the supplied source HTML.
 
 
     
 Local data 
Saved targets, named lists, and recent check summaries remain only in this browser.
 
   
 
      
 
How it works
 
The parser scopes title, description, canonical, robots, hreflang, and JSON-LD evidence to the document head. Commented tags, script strings, and body-level decoys are ignored. HTML entities are decoded, JSON-LD object keys are normalized structurally, and heading levels plus text are compared as an ordered outline.
 
Limits
 
The result covers only the two supplied documents. It does not fetch a live URL, render JavaScript, inspect HTTP headers, read robots.txt, verify hreflang reciprocity, validate schema eligibility, or determine whether a release should proceed. Malformed or unavailable evidence remains not evaluated.
 
Frequently asked questions
 
Does this fetch either page?
No. Supply each HTML document by pasting it or choosing a local .html file. Both documents and the comparison stay in this browser; neither page nor its contents are fetched or uploaded. A bounded anonymous usage ping may be sent after a run, but it contains no supplied HTML, URL, domain, IP, or identifier.
What happens when a signal cannot be evaluated?
The row stays not evaluated. A missing or malformed head, malformed JSON-LD, and crawler posture that HTML cannot establish are not treated as matches.
What counts as a regression?
The conservative regression label is reserved for removed observed evidence, removed hreflang declarations, a shorter heading outline or lost H1, and newly restrictive robots or crawler directives. Other differences are labeled changed for human review.
Why is AI-crawler posture not evaluated?
A pasted HTML document does not prove the site robots.txt policy or edge behavior for AI crawlers. The row is retained to make that evidence boundary explicit.
What does an export contain?
Exports contain only the normalized comparison rows and their states. They never include either pasted HTML document.
 
 
People also explore
ToolStaging vs. Production SEO DiffToolSEO Opportunity FinderToolPage and Website Authority CheckerArticleJoomla SEOArticleHugo SEOArticleURL Inspection Tool
 
 
 
Feature requests for Prepublish Comparison
 

Upvote what you want most. New ideas can be submitted from the floating Feedback menu;
      requests appear here once approved, and the most-wanted rise to the top.

   
Loading…
  
  ➕ Request a feature   
New requests are reviewed before they appear here.
 
     
  
  
 
 
 
 
     
     
Sobre a ferramenta
Grátis, sem cadastro. Compare um draft HTML exportar com o atual ativo HTML você fornecer antes approving um release.
Texto: Compare draft e ativo HTML locally across titles, descriptions, canonicals, robots directives, hreflang, heading outlines, e structured dados.
Recursos
  • O conservative regression rótulo é reserved para removido observado evidência, removido hreflang declarations, um shorter heading outline ou lost H1, e newly restrictive robots ou rastreador directives. Outro differences são labeled changed para humano revisão.
  • O parser escopos título, descrição, canonical, robots, hreflang, e JSON-LD evidência para o documentar head. Commented tags, script strings, e body-level decoys são ignored. HTML entities são decoded, JSON-LD object keys são normalized structurally, e heading levels plus texto são comparado como um ordered outline.
  • Exporte ou visualize o HTML completo do rascunho e da página ativa atual. Cole os dois documentos ou escolha dois arquivos .html locais. Entradas acima de 1 MiB são rejeitadas. Compare as linhas de sinais normalizados. Investigue primeiro as regressões e confirme se cada linha alterada é intencional. Baixe as linhas em CSV ou JSON para uma lista de verificação de lançamento. As exportações omitem o HTML de origem fornecido.
  • Cada linha relatórios correspondência, changed, regression, ou não avaliado. Um regression é um conservative signal-removal ou restriction rule—não um prediction de rankings, tráfego, ou release impacto. Revisão pretendido alterações em seus publicação contexto.
Como funciona
O parser escopos título, descrição, canonical, robots, hreflang, e JSON-LD evidência para o documentar head. Commented tags, script strings, e body-level decoys são ignored. HTML entities são decoded, JSON-LD object keys são normalized structurally, e heading levels plus texto são comparado como um ordered outline.
Limitações
  • Um pasted HTML documentar faz não prove o site robots.txt policy ou edge comportamento para AI rastreadores. O linha é retained para make que evidência boundary explicit.
  • O resultado covers somente o dois fornecido documents. Isso faz não buscar um ativo URL, render JavaScript, inspecionar HTTP cabeçalhos, leitura robots.txt, verifique hreflang reciprocity, validar esquema eligibility, ou determine whether um release deve proceed. Malformed ou indisponível evidência permanece não avaliado.
Perguntas frequentes
O que acontece quando um signal cannot ser avaliado?
O linha stays não avaliado. Um ausente ou malformed head, malformed JSON-LD, e rastreador posture que HTML cannot establish são não treated como corresponde.
Por que é AI-crawler posture não avaliado?
Um pasted HTML documentar faz não prove o site robots.txt policy ou edge comportamento para AI rastreadores. O linha é retained para make que evidência boundary explicit.
Faz este buscar either página?
Não. Fornecer cada HTML documentar por pasting isso ou choosing um local .html file. Both documents e o comparação stay em este navegador; neither página nor seu contents são buscado ou uploaded. Um limitado anonymous usage ping pode ser sent depois um run, but isso contains Não fornecido HTML, URL, domain, IP, ou identifier.
O que counts como um regression?
O conservative regression rótulo é reserved para removido observado evidência, removido hreflang declarations, um shorter heading outline ou lost H1, e newly restrictive robots ou rastreador directives. Outro differences são labeled changed para humano revisão.
O que faz um exportar contain?
Exports contain somente o normalized comparação linhas e seus states. They never include either pasted HTML documentar.