# Vectorize — bitmap to SVG tracing API

> Convert PNG/JPEG/WebP/BMP/TIFF images into compact, high-fidelity SVG.
> Free web UI at https://vectorize.erpfit.com. An ERPFit product.

## API

POST https://vectorize.erpfit.com/api/vectorize
  multipart/form-data fields:
    image     (file, required) — png/jpeg/webp/bmp/tiff, max 10MB
    settings  (JSON string, optional) — see below
  optional header: X-API-Key (dedicated processing lane + usage tracking)
  response: {"svg": "<svg…>", "engine": "tracecraft", "originalSize": n, "svgSize": n}
  errors: 400 invalid input, 401 bad API key, 429 queue full (Retry-After header), 503 timeout

settings JSON:
  {"preset": "auto"}                      — auto | logo | photo | poster | lineart (default auto)
  {"engine": "auto"}                      — auto | tracecraft | tracecraft2 | potrace2 | vtracer | potrace
  {"preprocessing": {"trim": false, "removeBackground": false, "sharpen": false}}
  {"tracecraft": {"maxColors": 16, "minRegion": 25, "tolerance": 0.5,
                  "drawStyle": "fill|strokeShapes|strokeEdges",
                  "shapeStacking": "cutouts|stacked", "groupBy": "none|color",
                  "gapFiller": "auto|on|off", "strokeWidth": 1}}

presets: auto (detect engine+settings from the image — recommended),
  logo (flat logos/icons), photo (photographs), poster (bold flat shapes),
  lineart (sketches/monochrome).

engines: tracecraft (in-house, beats vectorizer.ai on SSIM at ~2.7x fewer
  bytes on its own benchmark corpus), tracecraft2 (primitive recognition:
  rect/circle/ellipse → smallest files), potrace2 (color potrace),
  vtracer, potrace.

other endpoints:
  GET /api/health  — {"status":"ok", lanes, workers}
  GET /api/schema  — full parameter schema with ranges + defaults
  GET /docs/api    — human-readable API docs

example:
  curl -X POST https://vectorize.erpfit.com/api/vectorize \
    -F "image=@logo.png" -F 'settings={"preset":"auto"}'