Labs · Vectorizer
Pixels in, vectors out.
Zoom in — no blur.
Trace PNG, JPG or WebP into clean SVG with the open-source VTracer engine, compiled to WebAssembly and running entirely in your browser. Free, unlimited, nothing uploads.
Drop an image here, or click to choose
.png / .jpg / .webp (processed locally, never uploaded)
Engine defaults — the balanced starting point for flat-color art.
Advanced parameters
Everything runs in your browser — the image never uploads, and the page keeps working offline once loaded.
Where this tool comes from
The engine is visioncortex's open-source VTracer 0.6.5 — a Rust raster-to-vector tracer (6,400+ GitHub stars, MIT license) compiled to a ~140 KB WebAssembly module that runs in a Web Worker, so the page never freezes and the image never leaves your device. License and provenance are published at vtracer.LICENSE.txt; the engine repository is github.com/visioncortex/vtracer. Every SVG it produces carries a "Generator: visioncortex VTracer 0.6.5" comment in its header, so the output is self-identifying.
The five panel parameters (color precision, despeckle, corner threshold, curve fitting, layering) map one-to-one onto the engine's CLI parameters; the four presets were tuned against logo, illustration and photo test sets in July 2026. The WebAssembly file loads only on your first trace and is cached by the browser afterwards. Images longer than 2,048px on the long edge are downsampled before tracing — the page states both the source and the traced size, and since the engine fits shapes rather than pixels, this barely changes the result while running several times faster.
What it can't do
Photo vectorization here is stylization, not reconstruction: a geometric tracer slices continuous tones into flat patches, the patch borders are visible, and the file is often larger than the original. If you want photo-grade vector reconstruction, the paid leaders (vectorizer.ai and friends) run server-side AI shape inference and genuinely do better on photographs — at the cost of uploading your file and a monthly fee. On flat-color input — logos, icons, illustrations, stamps — the two routes converge, and a geometric engine is simply the right tool.
It also won't cut out subjects or produce semantic layers (every path is just a color patch), gradients become banded (gradient reconstruction is outside the engine's scope), and it needs a modern browser from 2023 or later (WebAssembly, module workers, OffscreenCanvas). On older browsers we say so plainly instead of degrading to a lower-quality fallback.
FAQ
Is my image uploaded anywhere?
No. Decoding, downsampling and tracing all happen in a Web Worker inside your browser — the page keeps working offline once loaded. Our server only delivers this page and the engine file itself.
How is this different from paid tools like vectorizer.ai?
Two different technical routes. This tool is classic geometric tracing (the open-source VTracer engine): free, unlimited, no watermark, and your file never leaves the device. The paid leaders run server-side AI models that reconstruct shapes — on photographs they genuinely do better. For logos, icons, illustrations and line art, geometric tracing is the right tool and the gap disappears; for photo-realistic reconstruction, that's their turf, and we say so.
What kind of image converts best?
Flat-color material: logos, icons, stickers, illustrations, colored comic panels, stamps and signatures (use the B&W preset for those). The rule of thumb is simple — the flatter the color patches and the cleaner the edges, the cleaner the vector. Photos and heavy gradients come out as posterized color blocks.
Why is the SVG bigger than my original image?
Vector file size scales with the number of shapes, not pixels. On photo-like input every small color patch becomes a path, and an SVG with thousands of paths easily outweighs a JPG. Drop color precision a notch or two, or use the Photo (stylized) preset, and the size falls by an order of magnitude. Logo-class input is the opposite — usually just a few KB.
Can I use the output commercially?
Your image, your call: the tool never touches your file, adds no watermark, and attaches no license terms of ours to the output. The engine itself is MIT-licensed open source. The one thing to check is the copyright of the source image — conversion doesn't change who owns the picture.