An SVG isn't pixels. It's a recipe for drawing.
A JPG or PNG stores the colour of every dot. An SVG stores instructions instead: a circle here, a curve there, this colour. Because it's text, you can read it, edit it, and draw it at any size without it going blurry. Change the code below and the picture follows.
Blow both up 8× and the difference is obvious.
The original badge, drawn once as SVG and once as a 64-pixel PNG. At the size it was made, they look identical. Enlarge them and the PNG runs out of dots, while the SVG is recalculated sharp at every size.
Perfect for logos. Hopeless for photos.
An SVG's size depends on how many shapes it has, not how big it's shown. A simple icon is a few hundred bytes; a photo traced into shapes becomes megabytes of code that still looks worse than a JPG.
Because it's code, an SVG can carry more than a picture.
SVG files can contain scripts and links to other files. That's why many sites refuse SVG uploads, and why email apps strip them out. We check for these parts in the SVGs you convert. The code at the top of the page is scanned live.
Open any SVG and see what it draws, what it carries and what it could run.
This page parses the file with your browser’s XML parser and walks every element. It finds scripts, event handlers, external loads and entity tricks, and sorts every byte by purpose. It also checks the size, viewBox and references, then runs a step-by-step optimiser and compares the result pixel by pixel. The file is only ever shown as an image, so nothing in it can run here. Everything happens on your device. .svgz files are decompressed first.
A drawing has no resolution until you pick one.
Every other image format on this site arrives with its dimensions already decided. A JPG is 4000 pixels wide because a sensor made it that way, and enlarging it invents detail that was never captured. An SVG has no pixels at all until something draws it, so the width and height you choose at conversion time are not a resizing step, they are the first time the image has had a size.
That is why the number you type matters more here than on any other converter. The sample file on this page declares viewBox="0 0 64 64" with a width and height of 64, but those attributes are a default, not a limit: the same file rendered at 1024 pixels is not upscaled, it is redrawn, and every curve is recalculated at the larger size. Ask for 64 pixels and you get a crisp small icon. Ask for 1024 and you get a crisp large one. Nothing in between is interpolated.
The decision that follows is which pixel format to land in, and it is mostly about transparency and edges. PNG and WebP keep the alpha channel, so a logo stays cut out against whatever sits behind it. JPG has no transparency at all and will fill those areas, usually with black or white, and its compression softens exactly the hard edges a vector drawing is made of. For a logo or an icon, PNG or WebP; for a photographic illustration that happens to be stored as SVG, JPG is defensible.
Two more things are lost in the move to pixels, and both are worth knowing before you convert rather than after. Styling goes: an SVG whose shapes use currentColor picks up the surrounding text colour and can be recoloured by CSS for hover states or dark mode, and a PNG of that same drawing is one fixed colour forever. Scripted or animated parts go too, which is usually a relief, because the inspector above exists to find exactly those scripts, event handlers and external loads before they reach a browser.
If the file is going to stay on the web and the target supports it, the best conversion is often no conversion: optimise the SVG and ship it. The optimiser on this page runs step by step and compares its output with the original pixel by pixel, so you can see that nothing shifted before you accept the smaller file. Rasterise when the destination cannot read SVG, which in practice means email clients, older document formats, and any upload form that rejects the extension outright.
An SVG is a document, which is why upload forms turn it away.
Every other image format on this page is inert data. A PNG is a grid of compressed pixels and a decoder; it cannot do anything. An SVG is markup, and markup can carry scripts, event handlers and references to other files. That single difference is why so many sites refuse SVG uploads outright and why email clients strip them, and it has nothing to do with the drawing itself.
The risk depends entirely on how the file is served. Placed in an ordinary image tag, a browser will not run scripts inside it. Embedded in an iframe, or opened directly from its own URL on your domain, that same file is a full document, and anything scripted in it runs with your origin's cookies. A form that accepts user-supplied SVG and serves it back from the main domain has handed visitors a way to run code as that site.
This is why the safe pattern is either to refuse SVG on upload or to serve it from a separate domain that holds no session, and why so many products choose the first option as the cheaper answer. If a form has rejected your logo and you were sure the file was fine, this is usually the reason: the rejection is about what the format can contain, not about what your particular file does contain.
The inspector above exists to tell those two cases apart. It parses the file with the browser's own XML parser, walks every element, and reports scripts, event handlers, external loads and entity tricks, sorting every byte by what it is for. A clean logo shows nothing but geometry and style, and you can say so with evidence rather than hoping.
When a destination will not take SVG and you cannot change that, rasterising is the fix, and it removes the question completely: a PNG or WebP of the same drawing cannot carry a script because the format has nowhere to put one. Convert at the size you need, keep the SVG as the editable master, and send the pixels where the markup is unwelcome.
SVG at a glance: the technical reference.
How to identify, serve and validate an SVG file, drawn from the W3C specification.
Every shape is a path, and a path is ten letters.
Inside the d attribute, each letter is a pen instruction followed by numbers. Capitals use absolute coordinates; lowercase ones move relative to the current point. Pick a command to see it drawn.
<path d="M 15 75 C 30 10, 70 10, 85 75"/>
From a W3C draft to every browser: the SVG timeline.
SVG was standardised early but took a decade to be usable everywhere, because Internet Explorer didn’t support it until version 9.
In and out of SVG.
Going to pixels is easy. Coming back from pixels means tracing, which works best on simple artwork.