diff --git a/README.md b/README.md
index 8b72a31..d00b79e 100644
--- a/README.md
+++ b/README.md
@@ -12,6 +12,7 @@ including DALL-E, OpenAI's LLMs, other LLMs api platform, also other image gener
 - **little tools**: base64 url to base64 data, base64 url to IMAGE, IMAGE to base64 url, regular llm text to word and "," only. etc.
 
 ## Update
+- **remove <think>...</think> tag for deepseek r1 via RegText_Node**
 - **Add use_jailbreak option for VisionLM api node**
   If your caption task rejected due to nsfw content, you can try to take on use_jailbreak.
   tested models: