Prompt Segments

README.md

@@ -1,12 +1,12 @@
 # FramePack Studio
 
-FramePack Studio is an enhanced version of the FramePack demo script, designed to create intricate video scenes with improved prompt adherence. This is very much a work in progress, expect some bugs and broken features. 
+FramePack Studio is an enhanced version of the FramePack demo script, designed to create intricate video scenes with improved prompt adherence. This is very much a work in progress, expect some bugs and broken features.
 ![screencapture-127-0-0-1-7860-2025-05-04-20_13_58](https://github.com/user-attachments/assets/8fcb90af-8c3f-47ca-8f23-61d9b59438ae)
 
-
 ## Current Features
 
 - **F1 and Original FramePack Models**: Run both in a single queue
+- **Dynamic Prompt Timeline**: Intuitive interface for creating time-based prompt changes
 - **Timestamped Prompts**: Define different prompts for specific time segments in your video
 - **Prompt Blending**: Define the blending time between timestamped prompts
 - **Basic LoRA Support**: Works with most (all?) hunyuan LoRAs
@@ -15,7 +15,6 @@ FramePack Studio is an enhanced version of the FramePack demo script, designed t
 - **I2V and T2V**: Works with or without an input image to allow for more flexibility when working with standard LoRAs
 - **Latent Image Options**: When using T2V you can generate based on a black, white, green screen or pure noise image
 
-
 ## Fresh Installation
 
 ### Prerequisites
@@ -28,6 +27,7 @@ FramePack Studio is an enhanced version of the FramePack demo script, designed t
 Install via the Pinokio community script "FP-Studio" or:
 
 1. Clone the repository:
+
    ```bash
    git clone https://github.com/colinurbs/FramePack-Studio.git
    cd FramePack-Studio
@@ -47,6 +47,7 @@ python studio.py
 ```
 
 Additional command line options:
+
 - `--share`: Create a public Gradio link to share your interface
 - `--server`: Specify the server address (default: 0.0.0.0)
 - `--port`: Specify a custom port
@@ -58,27 +59,39 @@ Add LoRAs to the /loras/ folder at the root of the installation. Select the LoRA
 
 NOTE: slow lora loading is a known issue
 
-## Working with Timestamped Prompts
+## Working with the Prompt Timeline
+
+The new dynamic prompt timeline interface makes it easy to create videos with changing prompts over time:
+
+1. **Add Prompt Segments**: Click the "+ Add Prompt Segment" button to add new timeline segments
+2. **Set Timing**: For each segment, enter the start time in seconds
+3. **Enter Prompts**: Type your prompt for each time segment
+4. **Remove Segments**: Click the ❌ button to remove unwanted segments (minimum one segment required)
+5. **Automatic Sorting**: Segments are automatically sorted by start time
+6. **Live Preview**: The timeline updates in real-time as you make changes
 
-You can create videos with changing prompts over time using the following syntax:
+The system will smoothly transition between prompts for a cohesive video. The interface automatically handles the conversion to the internal format required by the generation engine.
+
+### Legacy Format
+
+You can still use the traditional timestamp format directly if preferred:
 
 ```
-[0s: A serene forest with sunlight filtering through the trees ]
-[5s: A deer appears in the clearing ]
-[10s: The deer drinks from a small stream ]
+[0s: A serene forest with sunlight filtering through the trees]
+[5s: A deer appears in the clearing]
+[10s: The deer drinks from a small stream]
 ```
 
-Each timestamp defines when that prompt should start influencing the generation. The system will (hopefully) smoothly transition between prompts for a cohesive video.
+Each timestamp defines when that prompt should start influencing the generation.
 
 ## Credits
+
 Many thanks to [Lvmin Zhang](https://github.com/lllyasviel) for the absolutely amazing work on the original [FramePack](https://github.com/lllyasviel/FramePack) code!
 
 Thanks to [Rickard Edén](https://github.com/neph1) for the LoRA code and their general contributions to this growing FramePack scene!
 
 Thanks to everyone who has joined the Discord, reported a bug, sumbitted a PR or helped with testing!
 
-
-
     @article{zhang2025framepack,
         title={Packing Input Frame Contexts in Next-Frame Prediction Models for Video Generation},
         author={Lvmin Zhang and Maneesh Agrawala},