diff --git a/frontend/src/views/ArticleDetail.vue b/frontend/src/views/ArticleDetail.vue
index 9336800..484a03f 100644
--- a/frontend/src/views/ArticleDetail.vue
+++ b/frontend/src/views/ArticleDetail.vue
@@ -64,6 +64,64 @@ const publishedAt = computed(() => article.value?.published_at || article.value?
 const isOwner = computed(() => auth.isOwner)
 const categories = computed(() => (article.value?.category || '').split(',').filter(Boolean))
 
+/** 把"一坨"译文/原文按"中文句号"切成 <p> 段,改善"挤在一起"的观感。
+ *  优先按 \n 切(LLM 排版过的),没有换行再按句号/问号/感叹号切。
+ *  句中常见的"Mr./U.S."等缩写不会出现在中文译文里,按 6+ 字符才切,避免半句话被切。
+ */
+function splitIntoParagraphs(html: string): string {
+  // 已经是 HTML 标签(LLM 排版过),不动
+  if (/<(p|div|br)\b/i.test(html)) return html
+  // 按 \n 切
+  if (/\n/.test(html)) {
+    return html
+      .split(/\n+/)
+      .map((p) => p.trim())
+      .filter(Boolean)
+      .map((p) => `<p>${escapeHtml(p)}</p>`)
+      .join('')
+  }
+  // 一坨纯文字:按句号/问号/感叹号切,每段最多 140 字
+  const text = html.trim()
+  if (!text) return ''
+  const chunks: string[] = []
+  const re = /([。！？!?])(?=[^。！？!?]{0,140})/g
+  let last = 0
+  let m: RegExpExecArray | null
+  while ((m = re.exec(text)) !== null) {
+    const end = m.index + 1
+    if (end - last > 6) {
+      chunks.push(text.slice(last, end))
+      last = end
+    }
+  }
+  if (last < text.length) chunks.push(text.slice(last))
+  return chunks.map((p) => `<p>${escapeHtml(p.trim())}</p>`).join('')
+}
+
+function escapeHtml(s: string): string {
+  return s
+    .replace(/&/g, '&amp;')
+    .replace(/</g, '&lt;')
+    .replace(/>/g, '&gt;')
+}
+
+const translationBody = computed(() => {
+  const a = article.value
+  if (!a) return ''
+  if (a.body_zh_formatted) return a.body_zh_formatted
+  if (a.body_zh_html) return splitIntoParagraphs(a.body_zh_html)
+  if (a.body_zh_text) return splitIntoParagraphs(a.body_zh_text)
+  return ''
+})
+
+const originalBody = computed(() => {
+  const a = article.value
+  if (!a) return ''
+  if (a.body_html) return splitIntoParagraphs(a.body_html)
+  if (a.body_text) return splitIntoParagraphs(a.body_text)
+  return ''
+})
+
 async function rerunTranslation() {
   if (!article.value) return
   if (!confirm('重新翻译会消耗配额,确认?')) return
@@ -234,8 +292,7 @@ onMounted(load)
             <template #header>
               <span class="card-header-title">📖 文章译文(原始)</span>
             </template>
-            <div v-if="article.body_zh_html" v-html="article.body_zh_html" />
-            <div v-else-if="article.body_zh_text" class="article-body-fallback">{{ article.body_zh_text }}</div>
+            <div v-if="translationBody" class="article-body-fallback" v-html="translationBody" />
             <NText v-else :depth="3">暂无译文</NText>
           </NCard>
         </div>
@@ -254,8 +311,7 @@ onMounted(load)
             <template #header>
               <span class="card-header-title">📄 文章原文</span>
             </template>
-            <div v-if="article.body_html" v-html="article.body_html" class="article-body-fallback" />
-            <div v-else class="article-body-fallback">{{ article.body_text }}</div>
+            <div v-if="originalBody" class="article-body-fallback" v-html="originalBody" />
           </NCard>
         </div>
 
@@ -293,10 +349,28 @@ onMounted(load)
 
 .article-body-fallback {
   white-space: pre-wrap;
-  line-height: 1.85;
+  line-height: 1.95;
   color: var(--color-letter);
-  font-size: 16px;
+  font-size: 19px;
   font-family: var(--font-sans);
+  max-width: 720px;
+  margin: 0 auto;
+  letter-spacing: 0.02em;
+  text-align: justify;
+  text-justify: inter-ideograph;
+}
+
+/* 原始译文里的"中文句号"或"换行"在浏览器里经常被压成一坨,
+   强制段间加空行 + 段首缩进 2 字符,接近纸质书阅读感 */
+.article-body-fallback :deep(p),
+.article-body-fallback :deep(div) {
+  margin: 0 0 1.1em 0;
+  text-indent: 2em;
+}
+
+.article-body-fallback :deep(p:last-child),
+.article-body-fallback :deep(div:last-child) {
+  margin-bottom: 0;
 }
 
 .article-image {