LTX-2.5 (Pro): Text to Video
Generate video from text with LTX 2.5, a video model built for fast, high-quality production work. Type a prompt and get a cinematic clip with sharp detail, accurate motion, and support for native multishot scenes that hold character and style across cuts.
简介
This ComfyUI workflow turns a text prompt into a finished video clip using the LTX 2.5 model via the LtxApi25TextToVideo node. Your prompt (and optional multi-shot structure) is sent to the LTX 2.5 API, which returns generated video content with sharp detail and coherent motion. The downstream SaveVideo node then writes the result to your ComfyUI output folder as a ready-to-share file.
你需要提供
- 正在读取工作流
在 ComfyUI 中运行。 此工作流无法在本页面运行。你可以在 Comfy Cloud 中打开（已预先加载所需内容），或下载后在自己的 ComfyUI 中使用。
在 Cloud 中试用常见问答
The LtxApi25TextToVideo node calls the LTX 2.5 API, so you need an internet connection and valid API access. SaveVideo runs locally to write the final file.
Structure your prompt as multiple shots with clear, consistent descriptors for characters, wardrobe, lighting, and style. If your node version exposes a multi-shot field or shot list, use it; otherwise, separate shots within the prompt text. Reuse the same character and style cues across shots.
Adjust whatever controls your LtxApi25TextToVideo node exposes. Some versions include parameters like duration, aspect ratio/resolution hints, and a seed or randomness setting. If a control is not visible, it may not be supported by your build or API plan.
SaveVideo writes the rendered clip to your ComfyUI output folder using the filename or prefix you set, at the FPS you choose. The exact container/codec depends on your SaveVideo configuration.