<b>OmniLottie</b> is a versatile framework that generates high-quality vector animations from multi-modal instructions.
For flexible motion and visual content control, we focus on Lottie, a light-weight JSON formatting for both shapes and animation behaviors representation.
However, the raw Lottie JSON files contain extensive invariant structural metadata and formatting tokens, posing significant challenges for learning vector animation generation.
Therefore, we introduce a well-designed <b>Lottie tokenizer</b> that transforms JSON files into structured sequences of commands and parameters representing shapes, animation functions and control parameters.
Such tokenizer enables us to build OmniLottie upon pretrained vision–language models to follow multi-modal interleaved instructions and generate high-quality vector animations.
To further advance research in vector animation generation, we curate <b>MMLottie-2M</b>, a large-scale dataset of professionally designed vector animations paired with textual and visual annotations.
With extensive experiments, we validate that OmniLottie can produce vivid and semantically aligned vector animations that adhere closely to multi-modal human instructions.
OmniLottie is capable of generating high-quality vector animations across a wide spectrum of complexity — from simple icons to intricate animated illustrations.
It demonstrates remarkable versatility through multiple generation modalities, including <b>Text-to-Lottie</b>, <b>Text-Image-to-Lottie</b>, and <b>Video-to-Lottie</b>,
making it a powerful and flexible solution for diverse creative and professional animation tasks.
OmniLottie is built on a pre-trained vision-language model (Qwen2.5-VL) and incorporates a Lottie tokenizer.
The model tokenizes both text and image/video inputs as prefix tokens, while the Lottie tokenizer encodes vector animation commands into a unified representation space.
The Lottie tokenizer abstracts Lottie animations into compact sequences of animation commands and control parameters, removing redundant metadata while preserving full generative flexibility.
Our Lottie tokenizer converts raw JSON files into compact sequences of shape, effect, and animation commands with their associated parameters.
This representation substantially shortens sequence lengths (81% reduction compared to raw JSON), improves token efficiency, and enables OmniLottie to learn a clean distribution over animation primitives.
The tokenizer supports five fundamental layer types: Precomposition, Solid, Null, Shape, and Text layers, each identified by a unique type parameter.
<h2class="title is-4">Comparison with SOTA Methods</h2>
<h3class="title is-5">Text-to-Lottie</h3>
<divclass="content has-text-justified">
<p>OmniLottie achieves near-perfect success rates, the best FVD, and the strongest motion alignment compared to baselines including DeepSeek, GPT-4o, and Recraft.</p>
<p>OmniLottie ranks first in FVD, object alignment, and motion alignment while maintaining high reliability. Methods such as AniClipart and LiveSketch exhibit low success rates and significantly longer runtimes.</p>
The MMLottie-2M dataset consists of two million Lottie animations paired with multi-modal annotations including text descriptions, keyframe images, and rendered videos.
Data is collected from major online platforms (LottieFiles, IconScout, Flaticon, Iconfont, Icons8) and augmented with SVG-derived animations.
The dataset undergoes comprehensive processing including Lottie cleaning, spatial-temporal normalization, video rendering, and coarse-to-fine captioning using VLMs.
Our data processing pipeline includes five stages: (1) Data collection from web sources and SVG conversion,
(2) Lottie cleaning to remove non-visual and non-parameterizable layers,
(3) Spatial-temporal normalization to 512×512 canvas and unified timeline,
(4) Video rendering with random backgrounds, and
(5) Multi-modal annotation using a coarse-to-fine VLM captioning strategy.
</div>
</div>
</div>
<br>
<!-- Related Links -->
<divclass="columns is-centered">
<divclass="column is-full-width">
<h2class="title is-3">Related Links</h2>
<divclass="content has-text-justified">
<p>
<b>We thank the following excellent open-source works:</b>
</p>
<p>
<ahref="https://omnisvg.github.io/"><b>OmniSVG</b></a>: A unified framework for SVG generation that leverages pre-trained VLMs. We build upon its methodology for vector graphics generation.
</p>
<p>
<ahref="https://icon-shop.github.io/"><b>IconShop</b></a>: The first work leveraging LLMs to generate monochrome, icon-level SVGs. We referred to its parametric implementation.
</p>
<p>
<b>Concurrent Works:</b>
</p>
<p>
<ahref="https://aniclipart.github.io/"><b>AniClipart</b></a>: Animates static clipart using text-to-video diffusion priors.
</p>
<p>
<ahref="#"><b>LiveSketch</b></a>: Generates sketch animations through optimization-based approaches.
</p>
<p>
<ahref="https://starvector.github.io/"><b>StarVector</b></a>: Equips LLM with an image encoder for Image-to-SVG generation.
</p>
</div>
</div>
</div>
<!-- Acknowledgements -->
<divclass="columns is-centered">
<divclass="column is-full-width">
<h2class="title is-3">Acknowledgements</h2>
<divclass="content has-text-justified">
<p>
We thank all the contributors for dataset construction and valuable discussions.
</p>
</div>
</div>
</div>
</div>
</section>
<!-- BibTeX -->
<sectionclass="section"id="BibTeX">
<divclass="container is-max-desktop content">
<h2class="title">BibTeX</h2>
<pre><code>@article{omnilottie2025,
title={OmniLottie: A Unified Scalable Vector Animation Generation Model},
author={Author1 and Author2 and Author3 and Author4 and Author5 and Author6 and Author7 and Author8 and Author9 and Author10},
journal={arXiv preprint arXiv:XXXX.XXXXX},
year={2025}
}</code></pre>
</div>
</section>
<!-- Footer -->
<footerclass="footer">
<divclass="container">
<divclass="columns is-centered">
<divclass="column is-8">
<divclass="content">
<pstyle="text-align:center">
Source code mainly borrowed from <ahref="https://keunhong.com/">Keunhong Park</a>'s <ahref="https://nerfies.github.io/">Nerfies website</a>.
