CtrlK
Kaiber AI is a creative technology company that builds AI video generation tools for musicians, visual artists, and content creators. Founded in 2022 by Victor Wang and Eric Gao, the company is headquartered in Arcadia, California. Kaiber became widely known for producing AI-generated music videos for artists such as Kid Cudi and Linkin Park, and for its viral "Astral Jump" TikTok trend. The platform leverages open-source diffusion models and related AI frameworks to convert text, images, audio, and existing video footage into stylized animated content. As of 2025, Kaiber reported $2.6 million in annual revenue, a team of approximately 24 employees, and over 5 million user sign-ups.
Kaiber's founding story begins with Victor Wang and Eric Gao, childhood friends whose mothers immigrated to the United States from Shanghai together. Wang trained as a structural engineer at UC San Diego before earning a Juris Doctor in intellectual property law from the University of California College of the Law, San Francisco, in 2015. He worked as an IP and technology transactions associate at law firms including Goodwin and Orrick, Herrington & Sutcliffe.
Gao, known by the stage name oksami, pursued a career as a music producer and software engineer. He worked at Lockitron from 2014 to 2015, where he developed iOS features, and then served as a software engineer at Facebook from 2017 to 2019. On Spotify, oksami has accumulated over 53,000 listeners, and his YouTube channel has nearly 58,000 subscribers.
In February 2022, Wang and Gao launched their first startup, Secret Garden, a web3 music project built during the NFT boom. The venture's first sale grossed $1.2 million, with approximately $300,000 in profit. However, the collapse of FTX in November 2022 wiped out Secret Garden's funds and forced the company to shut down.
From the failure of Secret Garden, Wang and Gao pivoted to generative AI. The idea for Kaiber grew out of Gao's long-standing desire as a music producer to create visual content for his tracks without needing a large production budget or specialized video editing skills. In November 2022, Wang (CEO) and Gao (CTO) co-founded Kaiber alongside Jacky Lu (Head of Research) and Christine Zuniga (Head of Marketing).
The company was entirely bootstrapped from the start. Kaiber's technology built on several open-source AI projects, including AnimatedDiff, Automatic1111 (a popular Stable Diffusion web interface), ControlNet, and Deforum. On top of these foundations, Kaiber developed a proprietary layer that combined these tools into a single creative application.
Kaiber entered a beta period that attracted significant attention through two high-profile music industry collaborations. The company created AI-powered lyric videos for all 15 tracks on Kid Cudi's Entergalactic album, which accompanied a Netflix animated special of the same name released in September 2022. In February 2023, Kaiber served as the AI production studio for Linkin Park's "Lost" music video (discussed in more detail below).
During this beta period, the "Astral Jump" trend went viral on TikTok. Users uploaded videos of themselves jumping and used Kaiber's AI tools to transform the footage into animated sequences, often set to the song "The Enemy Between My Ears" by Sons Of Charlie. This organic virality drove massive sign-up growth.
By May 2023, when Kaiber officially exited beta, the platform had accumulated over 2 million users. In its first few months of operation, the company reported seven figures in gross revenue, a notable achievement for a bootstrapped startup.
Kaiber's public launch in May 2023 introduced several new features:
In November 2023, Kaiber launched a dedicated mobile app, expanding its reach to iOS and Android users. At the time of the mobile launch, the company announced partnerships with additional music artists and reported over 5 million total sign-ups. TechCrunch covered the launch, describing Kaiber as "the generative AI creative studio behind the music videos of popular artists Kid Cudi and Linkin Park."
On October 16, 2024, Kaiber launched Superstudio, a redesigned platform described as "an AI-native platform redefining how creatives interact with generative AI." Superstudio replaced the original Kaiber interface with an infinite canvas workspace that consolidated image generation, video creation, animation, and audio-reactive visuals into a unified environment.
Coinciding with the Superstudio launch, Kaiber announced its first external funding: a seed round led by EQT Ventures and Crush Ventures. The exact amount raised was not publicly disclosed. This marked a shift from the company's fully bootstrapped approach, providing additional capital for platform development and expansion.
Kaiber's technical foundation rests on several open-source machine learning projects, combined with proprietary integrations.
| Technology | Role |
|---|---|
| Stable Diffusion | Base image generation model using latent diffusion models |
| AnimatedDiff | Motion module for generating frame-consistent animations from text-to-image models |
| Automatic1111 | Web-based interface for Stable Diffusion, used as part of Kaiber's backend pipeline |
| ControlNet | Conditioning framework for controlling spatial composition, pose, and structure in generated outputs |
| Deforum | Animation framework that applies small frame-to-frame transformations using image-to-image diffusion |
Kaiber's proprietary layer fuses these components together, handling tasks like temporal consistency between frames, audio analysis for beat-synced visuals, and style transfer across video sequences.
With the launch of Superstudio in October 2024, Kaiber shifted to an aggregator model, integrating multiple third-party AI models alongside its own tools:
| Category | Integrated Models |
|---|---|
| Image generation | Flux (Black Forest Labs), Recraft, Stability AI, Topaz |
| Video generation | Luma AI, Veo (Google), Kling (Kuaishou), MiniMax, Mochi, Runway |
| Audio processing | Stability AI, AudioShake |
This multi-model approach allows users to choose the most suitable AI engine for their specific creative task rather than being locked into a single model's capabilities.
Kaiber's initial product suite offered four main generation modes:
Flipbook: A frame-by-frame animation tool that produces evolving, morphing visuals with a dreamlike quality. Each frame is generated individually, creating an effect where one image gradually melts into the next. Flipbook is well suited for abstract, psychedelic, or surrealist visual styles.
Motion: A tool designed to bring static images to life with smoother, more predictable movement. Users can upload a drawing or photograph and apply motion to specific elements. For example, a character illustration can be animated to walk, gesture, or look in different directions.
Transform: Kaiber's video-to-video tool. Users upload existing video footage and provide a text prompt describing a new artistic style. Transform then repaints the original video in the specified style while preserving the underlying motion and composition. This was the tool behind many of Kaiber's music video projects, including the Linkin Park collaboration.
Audioreactivity: A feature that synchronizes generated visuals to an uploaded audio track. Kaiber's AI analyzes the beat, tempo, rhythm, and energy of a song, then produces visuals that pulse, shift, and transform in time with the music. This tool made Kaiber particularly popular among musicians and DJs who needed visual accompaniments for their tracks.
Superstudio expanded and reorganized Kaiber's toolset around three core concepts:
Canvas: An open, infinite workspace where creators can generate, arrange, and refine visual assets using drag-and-drop functionality, adjustable sliders, and intuitive controls. Multiple canvases can be created for different projects.
Flows: Modular, customizable workflows that chain together multiple AI operations. Users can create repeatable creative pipelines, for example generating an image, applying a style, and then animating it in a single automated sequence.
Collections: An organizational system for managing generated assets, allowing users to group, tag, and retrieve their work efficiently.
Additional Superstudio capabilities include:
Kaiber created AI-powered lyric videos for all 15 tracks on Kid Cudi's Entergalactic album. The album accompanied a Netflix animated TV special of the same name, released on September 30, 2022. This project was one of Kaiber's earliest high-profile collaborations and helped establish the company's reputation in the music industry.
The music video for Linkin Park's "Lost" was one of the most prominent early examples of AI-generated content in mainstream music. "Lost" was a previously unreleased track from the Meteora sessions, included on the 20th Anniversary Edition of the album.
The video was directed by Maciej Kuciara and Emily Yang (also known as pplpleasr), co-founders of the web3 content studio Shibuya. Kaiber served as the AI production partner, with Jacky Lu (Kaiber co-founder and Head of Research) working as the principal AI artist on the project.
The production process involved several AI techniques:
The video was nominated for an MTV Video Music Award and received widespread media coverage as one of the first major music productions to feature AI video generation prominently.
Kaiber Labs developed custom AI models and visual outputs for Grimes' performance at Coachella 2024. The collaboration produced dynamic, futuristic visuals that accompanied the set. Kaiber's team worked from Grimes' creative moodboards to generate animations and visual elements tailored to her artistic direction.
Kaiber Labs co-produced animations for Yaeji's headlining set at Boiler Room in Brooklyn, New York, in collaboration with visual artist Weirdcore. The team created a series of imaginative animations and emoji-style characters for Yaeji's "Woofa-verse" visual universe, drawing inspiration from the artist's own moodboard.
Kaiber Labs partnered with Canadian artist Jon Rafman and Live From Earth, a Berlin-based techno collective, for an event at Academy LA on Hollywood Boulevard. In just 10 days, the team produced over 1,000 videos and 47 gigabytes of visual content mapped across multiple screens throughout the venue. The night featured sets from DJ Gigola, Two Shell, Brutalismus 3000, and others, with visuals morphing from 1980s game aesthetics to medieval-inspired dreamscapes.
Beyond these highlighted projects, several other well-known artists have used Kaiber's tools:
| Artist | Project |
|---|---|
| Wu-Tang Clan | AI-generated visual content |
| Don Diablo | Music video and visual production |
| Money Man | Visual content creation |
| Mike Shinoda | Creative work and collaboration with Kaiber |
Many of these artists discovered and adopted Kaiber organically rather than through formal partnerships.
Kaiber's pricing has evolved as the platform transitioned from its original product to Superstudio.
| Plan | Monthly Price | Annual Price | Credits Included | Key Features |
|---|---|---|---|---|
| Flex | Free | N/A | Pay-as-you-go only | 3 concurrent generations, 2 canvases, commercial use |
| Creator | $29/month | $279/year (~$23/month) | 1,400/month | 5 canvases, all models, custom LoRA training, 10-20% credit pack discounts |
| Pro | $149/month | $1,429/year (~$119/month) | 7,500/month | Unlimited canvases, all features, 20% credit pack discounts, priority access |
| Visionary | Custom | Custom | Custom | White-glove support, enterprise solutions |
Users on any plan can purchase additional credits:
| Pack | Credits | Price |
|---|---|---|
| Mini | 300 | $5 |
| Small | 1,000 | $15 |
| Big | 3,500 | $50 |
| Super | 20,000 | $250 |
Credits from subscriptions and purchased packs are pooled in the user's account. Different actions, video lengths, and AI models consume different amounts of credits.
Kaiber operates in the competitive AI video generation market alongside several other platforms:
| Platform | Developer | Strength |
|---|---|---|
| Runway | Runway AI | Professional video editing integration, consistent multi-shot generation |
| Pika Labs | Pika | Fast generation speeds, creative effects modules, affordable pricing |
| Sora | OpenAI | High visual quality, physics understanding, long-form generation |
| Kling | Kuaishou | Strong motion quality and video length capabilities |
| Luma AI | Luma AI | Dream Machine model, 3D-aware generation |
| HeyGen | HeyGen | Avatar-based video creation and localization |
| Synthesia | Synthesia | Enterprise-focused AI presenter videos |
Kaiber differentiates itself from these competitors primarily through its focus on music-driven workflows. The platform's audioreactivity features, beat sync tools, and deep ties to the music industry set it apart from general-purpose AI video generation tools. Its Superstudio platform also takes a multi-model aggregation approach, integrating outputs from Runway, Luma AI, Kling, and others rather than relying solely on a proprietary model.
| Detail | Information |
|---|---|
| Founded | November 2022 |
| Headquarters | Arcadia, California, United States |
| Founders | Victor Wang (CEO), Eric Gao (CTO), Jacky Lu (Head of Research), Christine Zuniga (Head of Marketing) |
| Employees | ~24 (as of 2025) |
| Funding | Seed round (October 2024), led by EQT Ventures and Crush Ventures; amount undisclosed |
| Revenue | $2.6 million (2025) |
| Users | 5+ million sign-ups (as of late 2023) |
| Website | kaiber.ai |