OmniSVG

OmniSVG suits designers and developers who want to generate or convert graphics into editable SVG, especially those comfortable working with vector code or running a local model. Its no-account demo makes occasional use straightforward, but the daily cap and hardware demands make it a poor fit for high-volume production.

Overview

OmniSVG turns natural-language descriptions into scalable vector graphics and converts raster images into editable SVG. You can preview results, download them, or copy the SVG code into a compatible editor. The converter accepts PNG, JPG, BMP, and WebP files up to 5 MB, a useful range for small assets but a firm ceiling for larger source images.

The free browser demo requires no account and shows how many generations remain that day. That makes it easy to try without a signup, though five generations per day is restrictive if you need to iterate on multiple designs or convert a batch of images.

Key features

Text, image, and character-reference generation

The project supports Text-to-SVG, Image-to-SVG, and Character-Reference SVG generation. Text prompts can produce new vector graphics, while image conversion offers a route from supported raster files to editable output. Character-reference generation broadens the project beyond those two workflows, though the browser demo's stated daily allowance limits how much experimentation is practical.

Editable SVG and local deployment

Downloadable SVG and copyable code give users a way to continue editing outside OmniSVG rather than treating the result as a finished bitmap. For users who want to run the project locally, the repository provides Gradio deployment instructions based on Python 3.10, Cairo, and dependencies tested with CUDA 12.1. That flexibility comes with substantial setup and compute requirements: the listed model files range from 7.69 GB to 17.2 GB, and reported GPU memory use reaches 26 GB for OmniSVG1.1_8B and 17 GB for OmniSVG1.1_4B.

The research project combines a pretrained Qwen-VL vision-language model with an SVG tokenizer and introduces MMSVG-2M, a dataset of two million annotated SVG assets. Those foundations explain the project's focus on vector generation, but they do not remove the practical limits of the demo or the resources needed for local inference. The code is licensed under Apache License 2.0, and a community-contributed ComfyUI plugin is also listed.

Pricing

Free — free. The free demo includes five generations per day, image uploads up to 5 MB, and requires no account. There is no paid tier to compare: occasional users can try the core workflows without committing, while anyone who needs more daily volume will need to consider another tool or a local setup.

Platforms

OmniSVG is available on the web, with Linux and macOS listed alongside self-hosted use. The local route is aimed at users prepared to install Python, Cairo, and CUDA-tested dependencies and manage large model files; the browser demo is the more practical option for a quick trial.

Who it's for

Web developers, UI/UX designers, content creators, and people making print or merchandise assets are the project's stated audiences. It is most compelling for someone who needs editable SVG from a prompt or a small raster image and can finish the work in a separate SVG-compatible tool. It is less suitable for users who need generous daily output, larger image uploads, or a turnkey local installation on modest hardware.

Pros and cons

• Pros: Text generation and raster conversion both produce editable SVG, giving users two ways to start from a graphic idea.
• Pros: No-account browser access and five free daily generations let occasional users try the workflow without signup.
• Pros: Self-hosted deployment and Apache 2.0 code licensing offer a local route for technically equipped users.
• Cons: Five daily generations and a 5 MB upload limit constrain repeated iteration and larger source images.
• Cons: Model downloads up to 17.2 GB and reported GPU memory needs up to 26 GB put local use beyond many everyday computers.
• Cons: SVG output still needs refinement in a compatible editor when the result is not ready to use as-is.

Alternatives

For a broader shortlist, browse AI Vector Generators.

• AnnoMotion is another freemium option, with web, Windows, and API platforms.
• Wix Logo Maker is worth considering when you need logo creation across web, Android, and iOS rather than general SVG generation.
• CleanVector offers web and API access, with a free plan that includes 10 credits a month and one permanent SVG download.
• Modulify AI Logo Maker is a web option for logo-focused work, with one free logo generation per day and a watermarked preview.
• SmartSVG may suit users who want a weekly credit allowance: its free plan gives 30 credits per week, including AI generation and SVG conversion.
• SVG.new is an alternative for vectorization, offering three free vectorizations per day and a built-in vector editor.
• SVGtoCode is another web-based option.
• PlusVector offers web and API access, with a free plan capped at five credits per month.

Verdict

Choose OmniSVG if you want no-account access to both prompt-based SVG generation and small-image conversion, or can run its models locally on suitable hardware. Its strongest reason to choose is editable vector output across those workflows; its clearest reason to look elsewhere is the combination of a five-generation daily cap and substantial local compute demands.