Kling family
# LTX 2.3 Pro High-fidelity AI video generation and editing suite. ## Overview LTX 2.3 Pro is a state-of-the-art generative video model designed to meet the rigorous demands of professional marketing teams and creative agencies. By bridging the gap between text-to-video synthesis and advanced video manipulation, LTX 2.3 Pro provides a versatile toolkit for creating high-quality, cinematic content. Whether you are starting from a blank canvas with a text prompt, animating static brand assets, or refining existing footage, this model offers the precision required for high-end digital advertising. What distinguishes LTX 2.3 Pro is its robust handling of complex creative workflows, including audio-to-video synchronization, temporal extension, and targeted retakes. With support for resolutions up to 4K and frame rates reaching 50 FPS, the model ensures that the output meets broadcast and social media standards. Its ability to interpolate between frames and apply specific camera motions allows creators to maintain stylistic consistency across long-form video projects, making it an essential asset for rapid prototyping and final-cut production. ## Use Cases * **Social Media Ad Campaigns:** Generate high-engagement 16:9 or 9:16 video content from static product photography for Instagram and TikTok. * **Brand Storytelling:** Extend existing brand footage to create seamless transitions or longer narrative sequences for website headers. * **Dynamic Product Showcases:** Use the 'dolly' and 'jib' camera motion presets to add professional cinematic movement to static product shots. * **Audio-Driven Content:** Sync video generation to custom audio tracks, perfect for music-led marketing spots or rhythmic social media trends. * **Precision Editing:** Utilize the 'retake' functionality to swap out specific segments of a video without needing to regenerate the entire sequence. * **High-Resolution Assets:** Produce 4K-ready video assets for high-definition digital signage and large-format display advertising. ## Parameters LTX 2.3 Pro offers a comprehensive suite of input parameters to control every aspect of your video generation. Please refer to the configuration interface for the full list of selectable options, including frame rates, aspect ratios, camera motion effects, and task-specific settings like retake modes and duration controls. ## Tips for Best Results * **Start with High-Quality Assets:** When using image-to-video, ensure your source image is high-resolution and well-lit to improve the fidelity of the generated motion. * **Leverage Interpolation:** Use the 'last_frame_image' parameter to define the start and end points of a sequence, ensuring the AI maintains a consistent visual path. * **Mind the Resolution Constraints:** Remember that audio-to-video, retake, and extend tasks are optimized for 1080p; plan your workflow accordingly if you require 4K output for primary generation tasks. * **Experiment with Camera Motion:** Use 'dolly_in' or 'dolly_out' to create a sense of depth in static product shots, which significantly increases viewer retention. * **Iterative Retakes:** When using the retake feature, keep your 'retake_duration' tight to ensure the model focuses its computational power on the specific area needing adjustment. ## About LTX 2.3 Pro is provided by Lightricks and hosted via Replicate (Cog version 0.16.12). As part of the LTX model lineage, it represents a significant advancement in generative video technology, specifically engineered to provide creative professionals with a reliable, scalable, and high-performance solution for AI-assisted video production.
Model input reference derived from preset schema.
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| fps | select | No | 25 | Frame rate in frames per second |
| task | select | No | "image_to_video" | The generation task to perform |
| audio | audio | No | null | Input audio file for audio_to_video task |
| image | image | No | null | First frame image for image-to-video generation |
| video | video | No | null | Input video for retake or extend tasks |
| prompt | string | Yes | null | Text prompt describing the video to generate |
| duration | select | No | 6 | Duration of the video in seconds |
| resolution | select | No | "1080p" | Resolution quality of the generated video. Only 1080p is supported for audio_to_video, retake, and extend tasks. |
| extend_mode | select | No | "end" | Where to extend the video. Used for extend task. |
| retake_mode | select | No | "replace_audio_and_video" | What to replace in the retake section. Used for retake task. |
| aspect_ratio | select | No | "16:9" | Aspect ratio of the generated video |
| camera_motion | select | No | "none" | Camera motion effect to apply to the generated video. Use 'none' for no camera motion. |
| generate_audio | boolean | No | true | Generate audio for the video. Used for text_to_video and image_to_video tasks. |
| retake_duration | number | No | 2 | Duration in seconds of the section to edit. Must be at least 2 seconds. Used for retake task. |
| last_frame_image | image | No | null | Last frame image for image-to-video generation. When provided, the video interpolates between the first frame and this last frame. |
| retake_start_time | number | No | 0 | Start time in seconds of the section to edit. Used for retake task. |