Held Back for 20 Years by "Reskinned Software", This Handcrafted Animation Is Even More Appealing in the AI Era?
“?? Wait, is this really still the Live2D I know?”
Not long ago, Bilibili creator Kamirete published a Live2D work titled “Full-Body 3D Robot / Can Do Addition”, leaving many viewers, including myself, completely stunned.
With fluid motions, dynamic camera work, full 360-degree rotation, and support for facial and hand tracking... as a Live2D model, it even houses a built-in logic-gate addition calculator crafted entirely from scratch.
It is worth remembering that Live2D is inherently a technology rooted in 2D illustration and deformation logic, lacking true 3D spatial environments like 3D software. Yet this piece managed to achieve an effect astonishingly close to a real 3D model. Backed by formidable technical mastery and polish, Full-Body 3D Robot stood out among submissions worldwide at the 12th Live2D Creative Awards (Live2D Creative Awards), taking home the Model Grand Prize.
High-level 3D visual fidelity, dozen-second “micro-short animations”, versatile material textures... throughout this competition, astonishing works like this were everywhere to be found.
Celebrating its 20th anniversary this year, Live2D has long been ubiquitous across game dynamic character portraits, virtual live-streaming avatars, and more. From entertainment giants to indie developers, from streaming avatar rigs to live wallpapers, Live2D has established itself as an industry standard, commanding a market share of up to 95% in the VTuber avatar sphere.
What is even more fascinating is that around this creative tool, Live2D is fostering a distinctive creative ecosystem and media aesthetic. More and more creators are beginning to break free from the stereotypes of mere “avatar rigs”, “animated illustrations”, or “mini-animations”.
01 The Boundless Potential of “Mini-Animations”?
How is a Live2D animation actually created? In brief, the original artwork is first separated into cut-out layers with occluded areas redrawn, followed by rigging and binding mesh points onto the moving parts, and finally setting the parameters for motion and deformation.
Compared to 3D animation, this pipeline preserves the charm of hand-drawn 2D artwork far better without requiring the grueling frame-by-frame drawings of traditional 2D animation. It strikes an exquisite balance between expressive impact and cost efficiency, making it uniquely suited for VTuber avatars and game portraits that demand frequent “character presentation” with an anime character’s tangible sense of “breath” and “presence”.
These unique properties have also fueled Live2D’s grassroots creative ecosystem. Driven by accessible tutorials and the booming market for fan art and original character (OC) commissions, we can now see countless creators online sharing intricate, self-made Live2D creations. Many are no longer content with simple blinks and sways, actively striving toward cinematic visual storytelling and mood-building instead.
Another Chinese creator shortlisted in this year’s Live2D Creative Awards, Naoshuihui, showcased precisely this ambition in his short film “Invisible? Friends”.
“Held back” by VTuber software for 20 years, why does this handcrafted animation shine even brighter in the AI era?
Naoshuihui is a professional Live2D modeler specializing in VTuber rigs and live wallpapers, active on platforms like Bilibili, Douyin, and Xiaohongshu. However, he revealed that he originally started as an illustrator; it was only when making dynamic illustrations for his own art that he discovered his passion and talent for Live2D, subsequently diving headfirst into the craft.
The piece Invisible? Friends was not even created primarily for competition. Naoshuihui mentioned that the spark behind it was rather serendipitous.
He had been going through a rough patch, “feeling in a very low slump, both in mood and in career.” While browsing Xiaohongshu one day, he stumbled across an illustration of a little girl sitting on a swing: “Both the color palette and the expression resonated deeply with how I felt back then. The urge to turn her into Live2D welled up inside me uncontrollably.”
He promptly reached out, obtaining permission from character designer @Jiewang de Dajie 114514 and illustrator @Zhuanqian Yangjia Bibixingren, and began adapting the artwork into Live2D. Though an uncommissioned passion project, he recalled, “I poured all the stubborn resolve and frustration I had bottled up during that period straight into creating this work.”
The completed piece indeed possesses a singularly unique aura. The original illustration is adorable, lonely, and uncanny, and the Live2D adaptation magnifies these traits even further. Upon release, it surpassed 600,000 views on Bilibili, resonating with a vast audience.
Naoshuihui walked through his creative approach. When receiving an illustration, his goal is not simply to make every element move indiscriminately, but rather to analyze and stage the visual elements first—a process that closely mirrors directorial “staging” (enshutsu) in animation production.
“I singled out three pieces of visual information: the swing, the cat, and the shadow, and then experimented to see if these limited elements could be strung into a short narrative. Eventually, I decided to open the shot on the cat, having the camera track its movement to reveal the protagonist swinging.
(After that) I felt that just swinging back and forth was a bit monotonous, so I offset the foreground character and the background along the depth axis during the swing, which produced a powerful sense of space.
Next, the camera cuts back to the cat catching the scent of a friend. As the BGM hits its crescendo, a soft meow summons the hidden friends into view.”
On the animation front, the creator put in equal care. He mentioned closely observing real cats to ensure the standing, sitting, and walking postures felt natural: “For instance, cats usually close their eyes when they meow, and their noses are sharper than their eyes, so (in the work) it sniffs first to detect the friend’s presence.”
Another example is the longitudinal depth motion and the body swaying along with the swing, achieved via optical illusions using rotation deformers. Simulating front-to-back depth movement within an entirely flat 2D plane lends both kinetic realism and psychological uncanniness, conjuring a surreal “dreamcore” atmosphere.
While Naoshuihui focuses on mood and emotional expression, Kamirete, the creator of “Full-Body 3D Robot / Can Do Addition” mentioned at the beginning, ventured toward exploring the software’s limits and technical boundaries.
Kamirete started uploading original videos to Bilibili in 2013, having previously translated numerous drawing tutorials. Later shifting his focus toward Live2D, he began releasing extensive model showcases, behind-the-scenes breakdowns, and systematic instructional videos from 2023 onward, even translating official Live2D tutorial books. He has become one of the most authoritative Live2D content creators across the Chinese web.
Kamirete told us that he previously had no connection to the art industry at all; he first picked up Live2D simply to create dynamic illustrations for songs he had composed.
His path to mastering Live2D also diverged from most. Rather than relying on video tutorials, Kamirete preferred reading through the official software user manual entry by entry. While this made for a steeper learning curve, it gave him an understanding of the software’s underlying logic far deeper than average users. Later, by answering technical questions from other users online, he gradually grew into a professional Live2D creator.
From mesh generation and parameter deformation to SDK integration with Unity and other engines, all the way to implementing 3D effects and logic gates, even seasoned Live2D users are consistently amazed by how comprehensively and deeply he explores Live2D’s feature set.
However, Kamirete’s experiments pushing the software to its limits are never done just for the sake of novelty; they all stem from practical obstacles encountered during everyday creation, learning, and community exchanges.
For example, standard Live2D rigs typically rely on front-and-back layer stacking. But if one wants to create a floating accessory that freely rotates and flies around a character, the movement involves omnidirectional rotation—something traditional front-facing illustrations combined with simple deformations can hardly achieve.
This inspired Kamirete to attempt 3D modeling directly inside Live2D. Could one do the inverse of “2D-to-3D rendering”—“assuming a static camera, rotating the world and objects in 3D”? The various facets of the model essentially simulate “the projection of a 3D object onto a 2D plane.”
“Take a point on a cylinder’s surface: when the cylinder faces forward, if the point is at the dead center of the front view, then after a 90-degree turn, it must lie right at the edge of the front view. As long as every point maintains its correct position in the front projection, the visual sensation of a true 3D rotation will emerge.”
Similarly, the addition calculator logic gate feature in his work originated from his investigation into the “Input Normalization” function while poring over the official user manual. In Live2D’s documentation, “Input Normalization” was originally designed as an underlying parameter for physical simulations, such as making hair sway realistically in response to head turns.
Intrigued by the underlying math, Kamirete built a basic logic module capable of addition inside Live2D to help both himself and his followers understand it better, later integrating it into his robot model.
“Some of my creations that look like ‘scientific research’, including using Live2D for ‘3D modeling’ and ‘logic gates’, are ultimately aimed at crafting better Live2D models.”
02 Why Choose Live2D?
With so many animation tools on the market, why choose Live2D? How has Live2D, two decades after its inception, managed not only to secure widespread industrial adoption—but also continue attracting a massive community of independent creators and young fans, fostering a singular creative culture?
In our interviews, both Kamirete and Naoshuihui reiterated the same sentiment: the primal concept of “bringing original 2D art to life” has always been the beating heart of Live2D.
Looking back, this was indeed the founding vision behind Live2D. Around the turn of the century in Japan—especially within the gaming sector—the industry was enamored with the spatial freedom and fluid motion of 3D graphics, while traditional 2D hand-drawn art appeared increasingly sidelined.
Live2D’s original creator, Tetsuya Nakajo, saw an opportunity in this shift. His way of revitalizing 2D illustration was not to cling strictly to classical hand-drawn animation, but to pioneer a new workflow with software engineering: allowing exquisite illustrations to move while preserving their original visual fidelity, enabling artists to create moving images while retaining the joy of drawing.
For decades, “animation” had always been an industrial endeavor. Traditional frame-by-frame animation demands labor-intensive keyframes, in-betweens, and coloring; 3D animation requires an entirely distinct skillset and mindset. Live2D, on the other hand, allows illustrations to skip the grueling pipeline of traditional animation while still delivering stunning dynamic life.
Twenty years on, this philosophy aligns seamlessly with the needs of today’s “era of individual creators.”
In an era where anyone can create, many young creators aren't necessarily aiming to produce a full-length animated film. They might simply want to animate a favorite illustration or their own OC, or start their journey through lightweight formats like short-form video or music videos.
Live2D preserves the beauty and idiosyncratic style of hand-drawn art while providing an accessible, low-barrier way to bring movement to life, giving characters their own breath, gaze, and gestures.
Compared to other tools, Live2D offers a much smoother learning curve. For creators with an art background, the interface feels intuitive, preventing newcomers from feeling discouraged early on. This accessibility has sustained its enduring vitality.
Naoshuihui noted that recent feature updates—such as the Liquify tool and Blend Shape deformers—directly addressed the practical needs of Live2D artists, dramatically boosting production efficiency.
At the same time, the Live2D team has steadily refined workflows with mainstream creative software like Photoshop and After Effects. Examples include a Photoshop plugin that automatically detects and clears stray pixel artifacts before importing into Live2D, as well as an AE plugin that loads models and motion data directly inside After Effects without requiring image sequences.
Furthermore, the Chinese-speaking community has accumulated a vast library of tutorials, case studies, and behind-the-scenes breakdowns over the years. Official documentation and event announcements on the website have consistently received prompt, high-quality localization. On Bilibili, the official Live2D account actively reposts standout creations, collaborates with creators on instructional videos, and provides promotional support.
As a creator, Kamirete is particularly grateful for this support: “Especially for intricate videos—like full-pipeline layer separation or new version feature guides—production is immensely time-consuming. Without official backing, releases would take much longer, or I’d have to cut back on content.”
03 20 Years Since Inception: The Future of Live2D
According to a joint survey report by Live2D and Kadokawa, the market scale for games and applications utilizing Live2D technology has reached 35 billion RMB annually. Monthly active users of Live2D software (excluding plugins and SDKs) stand at 100,000 across the globe, with annual creator and freelance commission transaction volumes surpassing 300 million RMB.
With the rapid rise of AI, the barrier to generating video is dropping precipitously. Yet animation created with Live2D’s distinct “handcrafted feel” remains widely celebrated, enjoying an expansive market.
Whether in full-length animation, gaming character art, or short-form video, what truly moves an audience isn’t merely the fact that something “moves”—it is that the characters gain breath, weight, and presence, enriching the storytelling and inspiring lasting empathy, engagement, and support.
How a character pauses, how the camera pushes in, and how emotion is conveyed are all the deliberate fruit of human thought and labor. Unlike generative AI, Live2D does not erase the presence of the creative act itself; it firmly retains the creator’s motion design, camera framing, and aesthetic voice.
Whether it is Naoshuihui choreographing visual elements and motion curves, or Kamirete’s “scientific discoveries” born from probing software features and answering community questions, their works touch viewers and earn acclaim precisely because the unique fingerprints of authorial intent and expression remain unmistakable.
At the same time, Live2D’s medium-specific qualities may unlock fresh aesthetic possibilities for animation.
Traditional 2D animation is built on frame-by-frame linework, where movement revolves around character silhouettes. In Live2D, what continuously shifts is the model’s internal geometry; motion achieved through deformation and parallax frees creators from drawing-count limits. While fully preserving the textural brushwork of 2D art, the camera movements and animations can glide as smoothly as 3D.
In indie projects, hyper-smooth motion paired with subtle perspective distortion can generate an uncanny, “dreamcore”-like surrealism.
On the other hand, in high-standard commercial anime production, Live2D is showing immense promise, demonstrating unique production workflows.
In 2018, Live2D and Japanese anime powerhouse Aniplex announced a capital and business alliance, embarking on exploring the viability of producing full-length anime using Live2D.
Between 2019 and 2022, Live2D released a succession of animated short films—“Beyond Creation”, “Hero Beta”, and “TR3 STEP1”—with all character animation executed in Live2D. The goal was to integrate Live2D into standard anime production pipelines, attaining visual fidelity comparable to TV broadcast series and theatrical films while executing dynamic camera shots difficult to pull off by hand.
In 2024, Live2D collaborated with OLM to release the TV anime “Tono to Inu” (殿と犬), marking the first anime series animated almost entirely with Live2D. It faithfully preserved the original mangaka’s hand-drawn brushwork while using Live2D to vividly capture the squishy, buoyant antics of a corgi.
Meanwhile, mainstream Japanese productions such as Wind Breaker, Spy × Family, Belle, and Overman King Gainer have also begun incorporating Live2D techniques for intricate character close-ups and complex dynamic motions.
Looking ahead to January 2027, the flagship anime series “Successor of the Wind” (風を継ぐもの), co-produced by Aniplex, Live2D, and animation studio Drive, is set to air across 28 TV stations throughout Japan. As a period drama centering on the Shinsengumi, the series features intense action sequences, clashing swords, and complex ensemble acting—poised to become a defining benchmark for Live2D’s capabilities in commercial full-length anime.
2006–2026: From game art and VTuber avatars to fan shorts and commercial anime, Live2D’s core mission over the past two decades has never wavered: “bringing illustrations to life.”
Yet its industry applications, creative ecosystem, and visual aesthetics have broadened beyond measure: serving industrialized production demands, expressing the personal emotions of independent creators, and nurturing an animation language all its own.
Here’s to the next 20 years, and the many surprises Live2D has yet to bring.