Convert an image to SVG
Whatever you were handed — PNG, JPG, WebP, BMP or GIF — drop it here. The browser decodes it and the worker traces it into editable SVG paths.
- Input
- Image
- Output
- SVG
- Engine
- WASM, in your tab
- Price
- Free preview
Drop an image here, or paste
Your image is never uploaded. It is traced in this tab.Open the full converter
Preset
- 01Logo — set for this page16 colours
- 02Line artBlack & white
- 03Photo64 colours
- 04Embroidery8 colours · Cutout
- 05Vinyl4 colours · Cutout
- 06LaserBlack & white · Polygon
- Output · SVG · PDF · EPS [soon] · DXF [soon]
- Preview free · download with Pro €9.00 / month
- Your file stays in this tab
Why vectortrace
- Ticket 1 / 3
In-browser and private
The engine is WebAssembly running in a worker on your own machine. Your image is never uploaded — so there is nothing to leak, nothing to delete, and nothing to take on trust.
- Ticket 2 / 3
Editable paths, not a wrapper
The engine groups the image into colour regions, takes the boundary between them and fits curves to it. What comes out is real geometry — one fill per region, every node yours to move — produced in the tab and never on a server.
- Ticket 3 / 3
Formats print and production use
SVG and PDF today; EPS and DXF for cutters, engravers and embroidery software as the exporters land. Real files, opened in Illustrator, Inkscape and LibreCAD — not an SVG with a JPEG inside it.
Format facts
| Format facts | Image · in | SVG · out |
|---|---|---|
| Kind | Raster. Any of PNG, JPG, WebP, BMP or GIF. | Vector. Paths described in XML. |
| Scaling | Every raster format blurs or pixelates past its native size. | Sharp at any size. |
| Colour | Whatever the source carries; the tracer reads the decoded pixels. | One solid fill per path, in a palette you can edit by hand. |
| Editable in | Photoshop, GIMP, Affinity Photo. | Illustrator, Inkscape, Figma, Affinity Designer, a text editor. |
| Typical use | Whatever arrived in the brief — you rarely get to choose. | Web, icons, print artwork, the input to a cutter workflow. |
Worth knowing
The five raster formats differ in how they store pixels, not in what a tracer does with them: all five are decoded to RGBA by the browser before the worker sees anything. The practical difference is the noise each format adds. PNG, BMP and lossless WebP add none. JPG and lossy WebP add ringing near edges. GIF adds dithering.
That is why the controls that matter are the same everywhere: the palette size decides how many fills the result has, and the speckle filter decides how small a region has to be before it is not worth keeping. Presets are just sensible starting values for those two.
If you know what you are holding, the format-specific pages carry the specific advice — PNG for alpha, JPG for artefacts, BMP for scans, GIF for dithering.
How it works
Drop any raster image
PNG, JPG, WebP, BMP or GIF. The browser decodes it; the worker traces the decoded pixels.
Choose a preset
Logo for a flat mark, line art for a drawing or scan, photo for a photograph. The readouts tell you what changed.
Export
SVG and PDF today; EPS and DXF as those exporters land.
Benchmark
Measured, not marketed. The corpus, the method and every figure the run produced are published in full.
Methodology → /benchmarks- boundary error, p95
- 0.39 px
- region-count error
- 0.113
- node-count ratio
- 6.35×
Questions
- 01Which image formats can I convert to SVG?
- PNG, JPG and JPEG, WebP, BMP and GIF. Decoding is done by the browser, so the accepted list is the browser image list rather than something we implement separately.
- 02What is the largest image I can trace?
- In the browser, whatever your machine has memory for — tracing is local, so there is no server-side ceiling. The public API applies its own documented limits because that runs on our hardware.
- 03Is the image uploaded?
- No. The tracer is WebAssembly running in a Web Worker in this tab, so the pixels stay on your machine. There is no upload step, which is also why it keeps working with the network off.
Related
Related conversions