CLI Quick Start

Install

From source (Rust 1.82+)

cargo install --path .

Local development

cargo run -- convert examples/test_logo_benchmark.png out.svg

Commands

Command	Purpose
`convert`	Default single-image conversion
`logo`	Logo and icon oriented conversion (clean paths)
`premium`	Higher-fidelity conversion for richer images
`auto`	Let EdgeSVG choose the method
`smart`	Iterative search under quality and size constraints
`info`	Quick recommendation for an input
`analyze`	Raw image features and classification
`compare`	Quantify SVG quality against source raster
`render`	Produce a PNG preview from an SVG
`optimize`	Round coordinates and remove SVG noise
`benchmark`	Batch convert a raster directory
`benchmark-golden`	Re-render golden SVG corpus and compare

Examples

Convert with auto mode

edgesvg convert logo.png logo.svg --method auto --json

The --json flag outputs a machine-readable JSON report with quality metrics:

{
  "svg": "...",
  "requested_method": "Auto",
  "effective_method": "Logo",
  "fallback_from": null,
  "decision": "Detected as logo (monochrome, low unique colors)",
  "report": {
    "metrics": {
      "ssim": 0.9987,
      "edge_f1": 0.9423,
      "foreground_iou": 0.9812,
      "path_count": 3,
      "file_size": 1248
    }
  }
}

Logo conversion

edgesvg logo icon.png icon.svg --json

Premium high-fidelity

edgesvg premium illustration.png illustration.svg --target 0.98 --json

Inspect an image

# Quick recommendation
edgesvg info logo.png --json

# Raw image features
edgesvg analyze logo.png --json

Compare SVG quality

edgesvg compare logo.png logo.svg --json

Render SVG to PNG

edgesvg render logo.svg preview.png --width 512 --height 512

Optimize an SVG

edgesvg optimize logo.svg --precision 2 --json

Batch benchmark

edgesvg benchmark \
  --input-dir images/ \
  --output-dir results/ \
  --json-path results/report.json \
  --markdown-path results/report.md

Golden benchmark

Re-render golden SVG corpus into rasters and compare against the current engine:

edgesvg benchmark-golden --json

JSON Output

All commands support --json for machine-readable output. This is the recommended format for scripting:

# Pipe to jq for quick metric inspection
edgesvg convert logo.png logo.svg --method auto --json | jq '.report.metrics.ssim'

# Check if quality meets threshold
SSIM=$(edgesvg compare logo.png logo.svg --json | jq '.ssim')
if (( $(echo "$SSIM < 0.99" | bc -l) )); then
  echo "Quality regression detected"
  exit 1
fi

CI/CD Integration

Add quality gates to your build pipeline:

- name: Vectorize assets
  run: |
    edgesvg convert assets/logo.png dist/logo.svg --method auto --json > report.json
    SSIM=$(jq '.report.metrics.ssim' report.json)
    python3 -c "import sys; sys.exit(0 if float('$SSIM') >= 0.98 else 1)"

See the CI/CD guide for a complete example.