# OmniSVG/OmniSVG

[NeurIPS 2025] OmniSVG is the first family of end-to-end multimodal SVG generators that leverage pre-trained Vision-Language Models (VLMs), capable of generating complex and detailed SVGs, from simple icons to intricate anime characters.

Repository: https://github.com/OmniSVG/OmniSVG
Canonical: https://ross.abutalabs.com/products/omnisvg
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-03-01T14:47:29+00:00

## Health v2 (maintenance only)
Score: 51/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 70, release rhythm 35, longevity 36
- inputs: {"age_days": 513, "days_push": 185, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2590, forks 103 (observed 2026-08-28T04:07:02.895820+00:00)

## What it is
OmniSVG is a family of end-to-end multimodal SVG generation models built on pre-trained Vision-Language Models, released with inference code, model weights, training code, and the MMSVG/MMSVGBench datasets. It generates complex and detailed SVGs ranging from simple icons to intricate anime character illustrations.

## Use cases
- generate svg icons from text prompts
- convert images into editable vector graphics
- create detailed anime character svgs
- generate scalable vector illustrations with ai
- benchmark multimodal svg generation models
- train custom svg generation models

## When to choose
- you need ai-generated or image-derived SVG vector graphics
- you want a research-grade multimodal SVG generator with open weights and datasets
- you need a benchmark for evaluating SVG generation models

## When to avoid
- you need lightweight deterministic SVG conversion without GPU inference
- you only need simple raster-to-vector tracing rather than generative modeling
- you lack GPU resources for large multimodal model inference

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, data-generation
- domain: artificial-intelligence, machine-learning, graphics, image-processing
- platform: python
- tags: svg-generation, vision-language-model, multimodal, text-to-svg, image-to-svg, neurips-2025, vector-graphics, model-weights, benchmark, gpu, linux

## Member repositories
- OmniSVG/OmniSVG (main) score 51

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:02.895820+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:22:19.688082+00:00, confidence not recorded.
  - readme: https://github.com/OmniSVG/OmniSVG (fetched 2026-08-28T04:07:02.895820+00:00, sha d22a9a0deb00)
- Data as of 2026-08-30T08:39:29.467469+00:00.