diff --git a/PR_COMPRESSION.md b/PR_COMPRESSION.md
index f1b40c2b..101874c0 100644
--- a/PR_COMPRESSION.md
+++ b/PR_COMPRESSION.md
@@ -39,4 +39,4 @@ We use [tiktoken](https://github.com/openai/tiktoken) to tokenize the patches af
 4. If we haven't reached the max token length, add the `deleted files` to the prompt until the prompt reaches the max token length (hard stop), skip the rest of the patches.
 
 ### Example
-![](https://codium.ai/images/pr_agent/git_patch_logic.png)
+<kbd><img src=https://codium.ai/images/git_patch_logic.png width="768"></kbd>
diff --git a/README.md b/README.md
index 9138ae5c..10cb2fb0 100644
--- a/README.md
+++ b/README.md
@@ -2,8 +2,13 @@
 
 <div align="center">
 
-<img src="https://codium.ai/images/pr_agent/logo-dark.png#gh-dark-mode-only" width="330"/>
-<img src="https://codium.ai/images/pr_agent/logo-light.png#gh-light-mode-only" width="330"/><br/>
+
+<picture>
+  <source media="(prefers-color-scheme: dark)" srcset="https://codium.ai/images/pr_agent/logo-dark.png" width="330">
+  <source media="(prefers-color-scheme: light)" srcset="https://codium.ai/images/pr_agent/logo-light.png" width="330">
+  <img alt="logo">
+</picture>
+<br/>
 Making pull requests less painful with an AI agent
 </div>