1. This workflow is the LTX2.3 head and tail frame video workflow, offering multiple configurations to choose from (FP8 version, KJ version, GUFF version). The online version defaults to the FP8 version.

2. The best quality is still the original version. However, using the KJ master’s model (ltx 2.3 22b distilled_transformer_only_fp8_scaled) provides the optimal model balancing low configuration, speed, and high quality.

3. It is strongly recommended to use landscape mode. Portrait mode can generate videos, but the results are not as good.

4. The automatic prompt words have been tested multiple times and can also be used to generate video prompt words with prompt word templates in other agents.

5. If you need I2V image-to-video generation, turn on the switch.

6. Regarding resolution selection:

(1) The LTX output video dimensions must be multiples of 32. Therefore, to ensure a 16:9 video ratio, you can refer to the following dimensions: 512×288; 1024×576
(2) Other approximate 16:9 aspect ratios: 1920×1088
(3) Additionally, 1280×720/736 output videos are all 1280×704
(4) 1280×768 can output normally
(5) The higher the resolution, the better the effect, but the higher the configuration requirements

7. Default parameters (1280×736, 10 seconds) consume about 80 points per run and take 7 minutes

Get free generation coins: Click "Top right corner on your avatar" > Invitation code > Enter the invitation code to get 1000 free RH coins. Log in daily to get 100 coins!

Invitation code: rh v1443