The unified multimodal creation tool Keling O1 is launched, unlocking infinite creative possibilities.

Recently, the world's first unified multi-modal video and image creation tool, "Keling O1", was officially launched. Based on a brand - new video and image model, Keling O1 uses natural language as the semantic framework and combines multi-modal descriptions such as videos, images, and subjects. It integrates all generation and editing tasks into an all - in - one engine, building a new multi-modal creation workflow for users and achieving a one - stop closed - loop from inspiration to finished product.

A Unified Model to Solve All Video Creation Problems

As the first unified multi-modal video model, Keling O1 is based on the MVL (Multi-modal Visual Language) concept. It breaks the boundaries of traditional single - video generation task models and integrates various tasks such as reference - based video generation, text - to - video generation, first - and - last - frame - based video generation, video content addition and deletion, video modification and transformation, style redrawing, and shot extension into the same all - in - one engine. This allows users to complete the entire creative process from generation to modification in one stop without switching between multiple models and tools.

Relying on the deep semantic understanding ability of the Keling Video O1 model, the pictures, videos, subjects, and texts uploaded by users are all instructions in the eyes of Keling O1. The model breaks through modal limitations and can comprehensively understand a photo, a video, a subject, or even different perspectives of a character, accurately generating various details.

The multi-modal instruction input area of Keling O1 turns the cumbersome post - editing process into a simple conversation. Users don't need to manually create masks or set key frames. They only need to input instructions like "Remove passers - by", "Change daytime to dusk", or "Replace the protagonist's clothing", and the model can understand the video logic. It can automatically complete pixel - level semantic reconstruction, from local subject replacement to overall video style redrawing. In addition, it also supports functions such as picture/subject reference, instruction transformation (video content addition, deletion, shot type/angle switching, video modification tasks, etc.), video reference, first - and - last - frame input, and text - to - video generation.

To address the pain point of the poor consistency between characters and scenes in AI - generated videos, Keling O1 enhances the understanding of input images and videos at the underlying level. It can "remember" the protagonist, props, and scenes like a human director. No matter how the camera moves, the characteristics of the subject remain stable. Moreover, the model demonstrates strong multi - subject fusion ability. Users can freely combine multiple different subjects or mix subjects with reference pictures. Even in complex group scenes or interactive scenarios, the model can independently lock and maintain the characteristics of each character or prop, ensuring industrial - level feature consistency of the "protagonist" in different shots.

It is no longer limited to single - point tasks but supports "skill combinations". Users can ask Keling O1 to "Add a subject to the video while modifying the background" or "Modify the style simultaneously when generating based on picture references". This ability to generate multiple creative variations at once greatly expands the freedom of creation and makes creative chemical reactions possible.

Freely define the narrative duration to give each story its unique rhythm. Keling O1 returns the power to define time to creators, supporting free generation from 3 to 10 seconds. Whether it's a short visual impact or a long - drawn - out story presentation, it's all up to the users. It's worth noting that as part of the unified model, the first - and - last - frame function of Keling O1 will also support the selection of a 3 - to - 10 - second generation duration (coming soon), which will further enhance the narrative flexibility.

Also debuting is the Keling Image O1 model, which can achieve seamless connection from basic image generation to high - level detail editing. Users can generate images through pure text or upload up to 10 reference pictures for fusion and re - creation. This model has four core advantages: high feature retention to keep the main elements stable; accurate response to detail modifications to meet expectations; precise control of style and tone to maintain a unified picture atmosphere; and super - rich imagination to make creative presentations more powerful, truly achieving "what you think is what you get".

One Model for Multiple Video Creation Scenarios in Film, Self - Media, Advertising, and E - commerce

The new Keling O1 integrates generation and editing functions and can be widely applied to various scenarios such as film, self - media, advertising, and e - commerce. Whether it's creating a narrative from scratch or deeply reshaping existing materials, Keling O1 can flexibly call its reference and editing capabilities according to different needs and easily complete the creation.

In the field of film creation, with the super - consistent picture (subject) reference of Keling O1 and the subject library function, users can accurately lock the characters, costumes, and props of each shot and easily create multiple consecutive film shots. For video post - production and self - media creators, they can let Keling O1 automatically complete pixel - level intelligent repair and reconstruction by simply inputting prompt words like "Delete the passers - by in the background" or "Make the sky blue".

To solve the problems of high cost and long production cycle in traditional offline advertising shooting, users can quickly generate multiple cool product display ads by simply uploading product pictures, model pictures, and scene pictures and providing simple instruction descriptions, significantly reducing the cost of actual shooting. To address issues such as the trouble of booking models and the need for repeated shooting when changing backgrounds or costumes, Keling O1 can build a virtual runway that never ends. Users can upload model + clothing real - shot pictures and input instructions to perfectly restore the texture and details of clothing and mass - produce high - quality Lookbook videos.

It is reported that the powerful and comprehensive functions of Keling O1 are derived from in - depth innovation in the technical foundation. The new Keling Video O1 model breaks the functional fragmentation in video model generation, editing, and understanding and builds a new generative foundation. By integrating the Multimodal Transformer for multi-modal understanding and multi-modal long - context, it realizes the deep integration and unification of multiple tasks.