Kling

Kling is an artificial-intelligence content-generation brand developed by Kuaishou Technology, the Chinese internet company known internationally for its short-video and live-streaming services. The brand provides a web-based creative studio for generating videos and images from natural-language prompts, uploaded images, and other visual references. Its principal use cases include concept development, social-media content, advertising and marketing assets, storyboarding, visual experimentation, and other forms of short-form digital production. Kling became publicly visible in 2024, when Kuaishou introduced it as a text-to-video model and opened access beyond the company’s internal research environment. Its emergence placed the product in the rapidly expanding generative-video market alongside systems such as OpenAI’s Sora, Runway, Pika, Luma, and Google’s video-generation offerings. Kling’s early positioning emphasized relatively long generated clips, motion coherence, visual detail, and the ability to maintain recognizable subjects across a sequence. The service subsequently broadened from text-to-video generation toward image-to-video, reference-based generation, and image creation, presenting these capabilities through a unified consumer-facing interface. The product is associated with Kuaishou’s broader artificial-intelligence research and product-development activities. Kuaishou has strategic reasons to develop generative media tools: such systems can support creators on its video platforms, provide new forms of creative software, and extend the company’s expertise in recommendation, computer vision, multimodal understanding, and large-scale content infrastructure. Kling is therefore both a standalone creative application and a potential technology layer for Kuaishou’s wider media ecosystem. The brand’s service model has included account-based access and usage limits or credits, although specific availability, pricing, model versions, and access policies can change over time and may differ by market. The official product presentation describes Kling as a tool for making high-quality videos and images from text, images, and references. International availability has expanded through an English-language web presence, while the product also maintains a strong connection to the Chinese market and Kuaishou’s domestic technology organization. As with other generative-media products, Kling operates in a technically and legally sensitive field. Important ongoing issues include the provenance and licensing of training data, the creation of synthetic likenesses, misinformation, copyright, moderation, and the risk that generated material may be mistaken for documentary footage. Public information establishes Kling as an active Kuaishou product brand, but does not provide a consistently documented standalone corporate structure, independent financial reporting, or a separately identified founder. Its precise headquarters, founding team, and long-term commercial model should therefore not be treated as independently established facts. Kling AI is a generative artificial intelligence product developed and operated by Kuaishou, the Beijing-based Chinese technology company best known for its short-video platform. The service converts natural-language descriptions, or prompts, into generated video sequences and is positioned within the rapidly expanding text-to-video and AI-assisted content-creation market. Kuaishou introduced the first version of Kling AI in June 2024. Its initial public testing took place through KuaiYing, Kuaishou's video-editing application, connecting the model to an existing ecosystem of mobile video creation and distribution tools. The launch made Kling part of the global competition among generative-video services seeking to produce increasingly coherent scenes, subjects, camera movements, and transitions from short written instructions. The product's principal use case is prompt-based video generation. Users describe a scene or sequence, and the system attempts to synthesize moving images that reflect the requested subjects, settings, actions, and visual style. Kling's development has focused on motion quality and the handling of difficult visual events, including rapidly moving objects and substantial changes between scenes. It has also been presented as a tool for creators, marketers, and other users who want to produce visual material without conventional filming or animation workflows. Kling's underlying architecture combines a diffusion-based transformer, commonly described as a DiT system, with a self-developed three-dimensional variational autoencoder developed by Kuaishou. The 3D VAE is intended to compress spatial and temporal information together, helping the service retain video quality while improving the efficiency of model training. A full-attention spatiotemporal modeling mechanism is also used to capture relationships across frames and to represent motion and scene detail. Kuaishou expanded the product through successive model generations. Kling 1.6 arrived in December 2024 with improved video-generation capabilities. Kling 2.0 followed in April 2025, and Kling 2.1 was released in May 2025 with multiple quality modes. On June 5, 2025, the service marked its first anniversary with a referral program, illustrating Kuaishou's effort to broaden adoption beyond its original testing environment. The brand has also attracted criticism and security concerns. Its content policies are reported to reflect Chinese regulatory requirements, resulting in restrictions on prompts involving politically sensitive subjects, protests, or criticism of the Chinese government. In May 2025, security researchers identified a malware campaign using fraudulent Kling AI websites and advertisements to distribute information-stealing malware. The campaign was attributed to malicious impersonation of the service rather than to Kling AI itself, but it demonstrated the risks associated with the product's growing visibility. Kling AI remains an active Kuaishou product and represents the company's move from short-video distribution into generative media infrastructure. Its identity is closely tied to AI video creation, iterative model upgrades, and the broader effort to make complex video production accessible through natural-language interaction. Kuaishou introduced the first version of Kling in June 2024, initially making it available for public testing through KuaiYing, the company's video-editing application, which connected the new model to Kuaishou's established creator ecosystem. Its emergence placed the product in the rapidly expanding generative-video market alongside systems such as OpenAI's Sora, Runway, Pika, Luma, and Google's video-generation offerings. Kling's early positioning emphasized relatively long generated clips, motion coherence, visual detail, and the ability to maintain recognizable subjects across a sequence. The service subsequently broadened from text-to-video generation toward image-to-video, reference-based generation, and image creation, presenting these capabilities through a unified consumer-facing interface. Technically, Kling uses a diffusion-based transformer architecture (commonly described as a DiT system) enhanced by a self-developed three-dimensional variational autoencoder. The 3D VAE performs synchronized spatial and temporal compression, an approach intended to preserve important information across video frames while supporting more efficient training. The system also uses a computationally efficient full-attention mechanism as a spatiotemporal modeling component, designed to help the model represent motion, fast-moving subjects, fine detail, and sharp changes between scenes. Kuaishou expanded the product through successive model generations. Kling 1.6 was released in December 2024 with improved video-generation capabilities. Kling 2.0 followed in April 2025, and Kling 2.1 was released in May 2025 with different quality modes. On June 5, 2025, Kling introduced a referral program to celebrate the service's first anniversary and encourage further adoption. The brand's service model has included account-based access and usage limits or credits, although specific availability, pricing, model versions, and access policies can change over time and may vary by market. Its English-language web presence indicates an effort to serve users outside China, although Kuaishou remains its corporate parent and the product retains a substantial Chinese-market identity. Kling's development has not been free of controversy. Its moderation system has been criticized for applying strict restrictions consistent with Chinese regulatory requirements; prompts involving politically sensitive issues, protests, or criticism of the Chinese government may be blocked or constrained. Separately, in May 2025, malicious actors used fake Kling AI websites and advertisements to distribute information-stealing malware. The incident was a brand-impersonation campaign rather than an identified action by Kuaishou, but it reflected the security risks created by the popularity of generative-AI services. Publicly available information identifies Kling as a Kuaishou product rather than as a separately listed company. There is no reliable basis in the supplied reference material for naming an individual founder, specifying an independent headquarters, or reporting standalone revenue. Kling should consequently be documented as an active product brand within Kuaishou's technology portfolio, with its detailed organizational history and financial performance treated as requiring further verification.

Kling is a generative-artificial-intelligence product brand created by Kuaishou Technology. It emerged publicly in 2024 during a period when major technology companies and specialized startups were competing to make video generation practical for consumers, creators, and professional media teams. Kuaishou’s background in short-video distribution, live streaming, recommendation systems, computer vision, and large-scale media infrastructure provided a natural institutional setting for developing a video-generation product. The initial public identity of Kling centered on text-to-video generation. Users could describe a scene in natural language and receive a synthesized video, rather than assembling every visual element manually. Early public demonstrations and product descriptions highlighted cinematic-looking outputs, relatively extended clips, subject and scene continuity, and the generation of motion from textual instructions. These features were significant because maintaining temporal consistency and plausible movement is substantially more difficult for video models than producing a single still image. Kling’s development took place amid rapid advances in multimodal foundation models. During 2024, generative-video products moved from research demonstrations toward accessible web applications. Kling followed this transition by presenting its technology through a user-facing service rather than only as a research announcement. Access, usage allowances, model capabilities, and regional availability have evolved over time, so the exact feature set of any particular release should be checked against the official product documentation. The service subsequently developed into a broader creative studio. In addition to text prompts, the product has been described as supporting image-to-video workflows, visual references, and image generation. These capabilities allow users to begin with a still image or a desired visual identity and then request movement, transformation, or related scenes. Such workflows are useful for concept art, story development, advertising mockups, social content, and rapid visual prototyping. The brand’s English-language web presence also indicates an effort to serve users outside China, although Kuaishou remains its corporate parent and the product retains a substantial Chinese-market identity. Kling’s strategic importance extends beyond the application itself. For Kuaishou, generative media can support the company’s creator ecosystem, help automate parts of content production, and create new software products alongside its established social-video business. The same capabilities also create governance challenges. AI-generated video can imitate real people, reproduce recognizable artistic styles, create misleading events, or incorporate material whose copyright status is disputed. Questions about model training data, consent, labeling, moderation, and commercial use are therefore part of the brand’s operating environment. Publicly available information identifies Kling as a Kuaishou product rather than as a separately listed company. There is no reliable basis in the supplied reference material for naming an individual founder, specifying an independent headquarters, or reporting standalone revenue. Kling should consequently be documented as an active product brand within Kuaishou’s technology portfolio, with its detailed organizational history and financial performance treated as requiring further verification. Kling AI was created by Kuaishou, a Beijing-based technology company founded in 2011 and known for operating a major short-video platform. Kuaishou developed Kling as part of its expansion into generative artificial intelligence and AI-assisted media production. The service is designed to generate video from natural-language prompts, allowing users to describe desired scenes, subjects, actions, and visual details rather than producing every element through conventional filming or animation. The first version of Kling AI was made available for public testing in June 2024 through KuaiYing, Kuaishou's video-editing application. This initial distribution connected the new model to Kuaishou's established creator ecosystem and placed generative video alongside more conventional editing tools. The launch occurred during a period of rapid development in text-to-video systems, as technology companies competed to improve duration, motion consistency, visual detail, and the ability to interpret complex prompts. Kuaishou subsequently developed the product through several model generations. Kling 1.6 was released in December 2024 with improved video-generation performance. Kling 2.0 followed in April 2025, and Kling 2.1 was released in May 2025 with different quality modes. These updates show a product strategy based on frequent model iteration rather than a single static software release. On June 5, 2025, Kling introduced a referral program to celebrate the service's first anniversary and encourage further adoption. Technically, Kling AI uses a diffusion-based transformer architecture enhanced by a self-developed three-dimensional variational autoencoder. The 3D VAE performs synchronized spatial and temporal compression, an approach intended to preserve important information across video frames while supporting more efficient training. The system also uses a computationally efficient full-attention mechanism as a spatiotemporal modeling component. This is designed to help the model represent motion, fast-moving subjects, fine detail, and sharp changes between scenes. The service's development has not been free of controversy. Its moderation system has been criticized for applying strict restrictions consistent with Chinese regulatory requirements. Prompts involving politically sensitive issues, protests, or criticism of the Chinese government may be blocked or constrained. Separately, in May 2025, malicious actors used fake Kling AI websites and advertisements to distribute information-stealing malware. The incident was a brand-impersonation campaign rather than an identified action by Kuaishou, but it reflected the security risks created by the popularity of generative-AI services. Kling AI is an active product brand within Kuaishou's broader technology portfolio. Its principal identity is as a prompt-driven video-generation service, while its successive model releases and creator-oriented distribution reflect Kuaishou's attempt to combine short-video expertise with generative media technology. Kling is a generative-artificial-intelligence product brand created by Kuaishou Technology, a Beijing-based technology company founded in 2011 and known for operating a major short-video platform. Kuaishou developed Kling as part of its expansion into generative artificial intelligence and AI-assisted media production. The service is designed to generate video from natural-language prompts, allowing users to describe desired scenes, subjects, actions, and visual details rather than producing every element through conventional filming or animation. Kuaishou's background in short-video distribution, live streaming, recommendation systems, computer vision, and large-scale media infrastructure provided a natural institutional setting for developing a video-generation product. The first version of Kling was made available for public testing in June 2024 through KuaiYing, Kuaishou's video-editing application. This initial distribution connected the new model to Kuaishou's established creator ecosystem and placed generative video alongside more conventional editing tools. The launch occurred during a period of rapid development in text-to-video systems, as technology companies competed to improve duration, motion consistency, visual detail, and the ability to interpret complex prompts. Early public demonstrations and product descriptions highlighted cinematic-looking outputs, relatively extended clips, subject and scene continuity, and the generation of motion from textual instructions. These features were significant because maintaining temporal consistency and plausible movement is substantially more difficult for video models than producing a single still image.

• 2025

  Kling 2.0 is released

  [5]Kling AIen.wikipedia.org

  Kling 2.0 launches in April as the next major model generation.

• 2025

  Kling 2.1 adds quality modes

  [5]Kling AIen.wikipedia.org

  Kling 2.1 launches in May with multiple quality modes for users.

• 2025

  First-anniversary referral program

  [5]Kling AIen.wikipedia.org

  Kling introduces a referral program on June 5 to mark its first anniversary.

• 2024

  Public introduction of Kling

  [1]Kling AI official websiteklingai.com

  Kuaishou publicly introduced Kling as an AI video-generation model and product, entering the rapidly developing text-to-video market.

• 2024

  Kling AI enters public testing

  [5]Kling AIen.wikipedia.org

  The first version of Kling is launched for public testing in June through Kuaishou's KuaiYing video-editing application, entering the rapidly developing text-to-video market.

• 2024

  Kling 1.6 is released

  [5]Kling AIen.wikipedia.org

  Kuaishou releases Kling 1.6 in December with improved video-generation capabilities.

• 2011

  Kuaishou is founded

  [5]Kling AIen.wikipedia.org

  Kuaishou, the technology company that later developed Kling AI, is established in China.

• Expansion into a multimodal creative studio

  [1]Kling AI official websiteklingai.com

  The official product presentation broadened Kling's scope to include video and image generation based on text, images, and reference inputs.

A multimodal generative-media studio positioned for high-quality video and image creation, with emphasis on motion, subject consistency, and accessible creative production. A Kuaishou-backed generative video platform focused on turning natural-language prompts into increasingly detailed and coherent videos, with an emphasis on motion quality, scene changes, and creator-oriented production workflows.

• Generative AI application · 2024

  Kling AI

  Kling AI is a web-based generative-media service for creating videos and images from natural-language prompts, uploaded images, and reference materials. Its capabilities have included text-to-video, image-to-video, and related multimodal workflows. The product is designed for creators, marketers, designers, storytellers, and general users who want to produce visual material without conventional frame-by-frame animation or full video-production workflows. The system is designed to represent subjects, actions, motion, scene details, and transitions, and is supported by Kuaishou's diffusion-transformer architecture and 3D variational autoencoder technology. Features, model versions, usage limits, and regional access may change as the service develops. Kling AI is a prompt-based video-generation service that converts natural-language descriptions into synthetic video. It is intended for creators and other users who want to produce moving-image content without relying entirely on conventional filming, animation, or manual editing. The system is designed to represent subjects, actions, motion, scene details, and transitions, and is supported by Kuaishou's diffusion-transformer architecture and 3D variational autoencoder technology.

• Generative video model · 2024

  Kling 1.6

  Kling 1.6 is a December 2024 model generation that brought reported improvements to Kling AI's video-generation capabilities. It formed part of Kuaishou's continuing effort to improve the service after its initial public testing release.

• Generative video model · 2025

  Kling 2.0

  Kling 2.0 is the April 2025 generation of the Kling AI model family. It continued the product's iterative development in prompt-based video synthesis and followed the earlier Kling 1.6 release.

• Generative video model · 2025

  Kling 2.1

  Kling 2.1 is the May 2025 model generation. Its notable user-facing addition was the introduction of different quality modes, allowing users to select among output-quality options.

• 2025

  Fraudulent Kling AI sites used to distribute malware

  In May 2025, malicious actors used fake Kling AI websites and advertisements to distribute information-stealing malware. The incident involved impersonation of the brand and was not identified as an operation conducted by Kling AI or Kuaishou.

  Source

• Criticism of politically restrictive content moderation

  Kling AI has been criticized for content controls that align with Chinese government regulations. Reported restrictions affect prompts concerning politically sensitive topics, protests, and criticism of the Chinese government.

  Source

Read more

• 2025First-anniversary referral program

  International

  Kling launched a referral program on June 5, 2025, to celebrate one year since the service entered public testing and to promote user growth through referrals.

  Source

• 2025

  Release Kling 2.0 and Kling 2.1Generation change

  Background

  Kuaishou pursued rapid model iteration in the competitive generative-video market.

  Action taken

  Kling 2.0 was released in April 2025, followed by Kling 2.1 in May with different quality modes.

  Aftermath

  The releases expanded Kling's model family and gave users additional output-quality choices.

  Source

• 2024

  Launch of Kling as a public AI-video productProduct launch

  Background

  The generative-video market was moving from research demonstrations toward consumer-accessible applications, while Kuaishou had relevant expertise in visual media and creator platforms and sought to extend its short-video and editing ecosystem into generative media. Kuaishou sought to extend its short-video and editing ecosystem into generative media by making its new text-to-video service available to the public.

  Action taken

  Kuaishou released the first version of Kling for public testing through the KuaiYing video-editing application, introducing it as a text-to-video system and making the technology available through a user-facing product experience.

  Aftermath

  Kling became one of the better-known Chinese entries in the international AI-video market and later presented broader image and reference-based generation capabilities. The launch established Kling AI as Kuaishou's generative-video product and was followed by successive model generations. The launch established Kling as Kuaishou's generative-video product and one of the better-known Chinese entries in the international AI-video market; it was followed by successive model generations and later broader image and reference-based generation capabilities.

  Source

• 2024

  Release Kling 1.6Generation change

  Background

  Following the initial public-testing release, Kuaishou continued to improve the quality and capability of its video-generation system.

  Action taken

  Kuaishou released Kling 1.6 in December 2024 with improved video-generation capabilities.

  Source

• 2025

  Kling 2.0 expands the model generation[5]Kling AIen.wikipedia.org

  Product launch · Product generation

  Kuaishou launched Kling 2.0 in April 2025, continuing the service's sequence of model upgrades.

• 2025

  Kling 2.1 introduces quality modes[5]Kling AIen.wikipedia.org

  Product launch · Product generation

  Kling 2.1 was released in May 2025 and introduced different quality modes for users.

• 2025

  Kling AI launches anniversary referral program[5]Kling AIen.wikipedia.org

  Campaign · Other

  On June 5, 2025, Kling AI marked its first anniversary by introducing a referral program intended to encourage user acquisition and sharing.

• 2024

  Kuaishou introduced Kling as a text-to-video model[1]Kling AI official websiteklingai.com

  Product launch

  Kuaishou presented Kling as a generative video system capable of producing clips from written prompts. The launch made the company a visible participant in the emerging AI-video market.

• 2024

  Kuaishou launches public testing of Kling AI[5]Kling AIen.wikipedia.org

  Product launch

  Kuaishou introduced the first version of Kling AI in June 2024 through its KuaiYing video-editing application, opening the generative-video service to public testing.

• 2024

  Kling 1.6 released with improved video generation[5]Kling AIen.wikipedia.org

  Product launch · Product generation

  Kuaishou released Kling 1.6 in December 2024 as an updated version of the generative-video system, with reported improvements to video-generation capabilities.

• Kling expanded toward multimodal video and image creation[1]Kling AI official websiteklingai.com

  Product generation

  The official service description presents Kling as a studio that can generate videos and images from text, images, and reference materials, broadening the product beyond its initial text-to-video framing.

1. 1Kling AI official websiteklingai.com
2. 2Kling AI: Next-Gen AI Video & Image Generatorkling.ai
3. 3Kling AI Release Notes & Version History | Kling AIkling.ai
4. 4Kling AI Release Notes & Version History | Kling AIkling.ai
5. 5Kling AIen.wikipedia.org