Track Hyper | "Keling AI + Lovart": Integration of AI Toolchain

Author: Zhou Yuan / Wall Street News

Recently, KUAISHOU's Keling AI has reached a deep cooperation with the global design intelligence agency Lovart. The two parties have achieved technical collaboration through API interfaces, with Lovart embedding Keling AI's video generation capabilities into its design process, supporting the automatic generation of complete video content exceeding one minute for single tasks, covering various scenarios such as advertising marketing and Vlogs (Video Blog/Video Weblog).

This cooperation is akin to Schumpeter's concept of "creative destruction" in "The Theory of Economic Development"—when generative AI technology is deeply integrated with design processes, traditional content production models begin to undergo a qualitative change from tool iteration to ecological reconstruction; this cooperation represents an important advancement of generative AI technology in the field of content creation, indicating that the integration of AI toolchains has entered a new stage.

Integrating AI Toolchains to Adapt to Change

From an industrial background perspective, China's content creation market is undergoing structural transformation.

According to IDC's report on "China's Generative AI Application Development Platform Market," the market size of generative AI software in China is expected to reach USD 3.54 billion by 2025, with a continuous increase in demand for intelligent computing power.

As generative AI technology matures, industries such as advertising marketing and film production are experiencing a surge in demand for efficient content production, making traditional human-dependent creation models increasingly inadequate to meet the market's rapid iteration requirements.

The cooperation between Keling AI and Lovart is a direct response to this industrial trend.

Keling AI, as KUAISHOU's self-developed video generation large model, has its technological advantages mainly reflected in three aspects: generation capability, multimodal collaboration, and commercialization maturity.

In terms of generation capability, Keling AI supports 1080P HD video output, achieving a 95% improvement in image quality and dynamic quality through the v1.5 high-quality model, supporting video generation of up to 3 minutes, far exceeding Douyin's Jiemeng AI (5 seconds) and Tencent's Hunyuan (16 seconds for text-to-video, 5 seconds for image-to-video).

It must be clarified that the "3 minutes" of video length generated by Keling AI actually refers to the total length of multiple rewrites by the user of the initial 4-5 second video, rather than being completed in one go; the single-generation duration is still limited by the model architecture (for example, Keling 2.1 Master Edition has a single generation limit of 10 seconds), and the rewriting function relies on the coherence of prompts, with complex scenes potentially leading to content gaps.

Multimodal collaboration: By combining a Sora-like DiT structure with a Flow diffusion model, Keling AI achieves industry-leading levels in physical simulation (such as fluid motion and gravitational effects) and detail representation (such as hair and facial muscle dynamics).

On the commercialization front, as of June this year, Keling AI has provided API (Application Programming Interface) services to over 10,000 enterprise clients, covering fields such as advertising marketing and film animation, with a high renewal rate among enterprise clients.

Lovart is the world's first intelligent agent in the design field, with its core value lying in building a full-link automation capability from demand understanding to finished product output Through natural language interaction, Lovart's users only need to input commands like "30-second video advertisement, 6-frame storyboard," and Lovart can automatically complete tasks such as script writing, character design, and scene layout planning, while calling on Keling AI to generate video content.

For example, in the "VIVIDPIX" brand launch case, Lovart not only generated a 3D logo animation but also created a brand theme song in sync with Suno AI, achieving a preliminary unity of visual and auditory elements.

The core of this collaboration lies in the integration of the "AI toolchain."

Lovart deeply embeds the video generation process into the design workflow by calling Keling AI's API, addressing the shortcomings of traditional design tools in dynamic content production.

Taking the visual system design of a plant-based milk brand as an example, while Lovart calls Keling AI to generate dynamic posters, it automatically matches background music and optimizes lighting effects, reducing the entire process from demand to finished product from 72 hours to 45 minutes, achieving an efficiency increase of over 5 times.

This "design-generate-optimize" closed loop significantly lowers the professional creation threshold, making it particularly attractive to small and medium-sized enterprises, as design costs can be reduced by about 70%.

The Invisible Hand of Technology

This collaboration directly drives the transformation of content production from "labor-intensive" to "AI creative."

In the field of advertising and marketing, the combination of Keling AI and Lovart can achieve "generate advertisement videos with one sentence," for example, inputting "Classic Red, Define Your Power," the system automatically generates a storyboard video featuring close-ups of lipstick and party scenes, along with a magnetic female voiceover.

As Kevin Kelly, author of "Out of Control," said, "The most profound technologies are those that disappear." When the AI toolchain is integrated into the production process, its value transcends the technology itself, becoming the underlying infrastructure for industry operations.

This capability enables companies to quickly respond to market trends, shortening the production cycle of marketing materials from weeks to hours.

For Keling AI, this collaboration further consolidates its advantages in the B-end market.

According to Kuaishou Technology's Q1 2025 financial report released on May 27, Keling AI's revenue scale has reached 150 million yuan, with nearly 70% of revenue contributed by P-end paid subscription members.

Through collaboration with Lovart, Kuaishou Technology's AI video technology advantages can extend into the design field, forming differentiated competitiveness; Lovart, by integrating Keling AI's video generation capabilities, also compensates for its shortcomings in dynamic content production.

Currently, Keling AI's main competitors in the video generation field are Tencent's Hunyuan and ByteDance (i.e., Dream AI). The former's video generation model can achieve 16-second text-to-video generation and supports multi-angle shot switching; Dream AI has launched a 2K resolution video generation function with a maximum length of 5 seconds.

Therefore, Keling AI needs to continuously optimize model performance, such as improving the accuracy of Chinese semantic understanding and strengthening cooperation with ISVs (Independent Software Vendors) to expand application scenario coverage; meanwhile, Lovart still faces generation efficiency issues during the internal testing phase: generating a 5-second video takes 5 minutes, and stability still needs improvement The collaboration between Keling AI and Lovart is essentially a handshake between technological rationality and creative sensibility.

This collaboration, a collision of "creativity and creativity," does not boast "epoch-making" grand statements, but it genuinely transforms video generation from a "privilege of the few" into a "tool for the many," driving the content industry towards a more efficient and inclusive direction through incremental technological improvements.

As Da Vinci wrote in "On Painting," "Art can express imagination more precisely with the help of science."

When AI technology becomes the "new brush" for creative expression, the industry witnesses not the technology surpassing human imagination, but rather human-machine collaboration opening up new possibilities: on the canvas woven by computing power, the scroll of creativity's ocean is gradually unfolding