diff --git a/apps/site/docs/en/index.mdx b/apps/site/docs/en/index.mdx
index 59698a56b..285f7c19f 100644
--- a/apps/site/docs/en/index.mdx
+++ b/apps/site/docs/en/index.mdx
@@ -62,7 +62,7 @@ Midscene wants to provide a way to make automation more stable and easier to deb
 What's more, there is a playground in the report file for you to adjust your prompt without re-running all your scripts.
 
 <p align="center">
-  <img src="/report.gif" alt="visualized report" />
+  <img src="/report.gif" alt="visualized report" loading="lazy" />
 </p>
 
 ## Support both general-purpose LLM and open-source model
diff --git a/apps/site/docs/zh/index.mdx b/apps/site/docs/zh/index.mdx
index e6d3ba94f..12c5d4bb2 100644
--- a/apps/site/docs/zh/index.mdx
+++ b/apps/site/docs/zh/index.mdx
@@ -48,10 +48,9 @@ Midscene 希望让自动化脚本变得更稳定、更易于调试，因此我
 此外，Midscene 报告里还集成了一个 Playground，用以在报告中重新运行 Prompt 并进行调优。
 
 <p align="center">
-  <img src="/report.gif" alt="visualized report" />
+  <img src="/report.gif" alt="visualized report" loading="lazy" />
 </p>
 
-
 ## 自定义模型
 
 Midscene 同时支持通用 LLM 和开源模型。你可以使用通用的 LLM 模型，如 `gpt-4o`，它在大多数场景都表现很好。