diff --git a/README.md b/README.md
index 289a877..ae33ec7 100644
--- a/README.md
+++ b/README.md
@@ -3,7 +3,7 @@
 This project is a demo for a websocket-based input synchronization method inspired by [GGPO](https://www.ggpo.net/).   
 ![screenshot-1](./screenshot-1.png)
 
-Please checkout [this demo video](https://pan.baidu.com/s/1ge9fEZNsRA8f6dpBKbdSjw?pwd=854g) to see whether the source codes are doing what you expect for synchronization.
+Please checkout [this demo video](https://pan.baidu.com/s/1aM6e8IWaJszFCYAsRjt19g?pwd=z02c) to see whether the source codes are doing what you expect for synchronization.
 
 The video mainly shows the following feature (yet I'm not surprised if they're not obvious): when a player didn't have its input arrived at the backend in time (e.g. due to local lag, network delay or reconnection), backend forces confirmation of a prediction of its own and sends the confirmed input together w/ a reference render frame to that player.