Recipes

Each recipe begins with a concrete image task and shows an actual input, generated outputs, complete code, and the current verification boundary. The four interfaces implement the same product intent without assuming identical defaults or encoded bytes.

Choose by task

I/O and conversion

Start with JPEG-to-WebP conversion or connect Reader/Writer APIs to streaming I/O. Format conversion · Streaming I/O

Geometry and canvas

Compare fit, cover, crop positions, resize, crop, rotate, and flip. Resize · Geometry basics

Composition and text

Add text or logo watermarks and compare Over, Multiply, SVG, and Pango text paths. Watermarks · Layer composition

Pixels and enhancement

Work with channels, alpha, brightness, saturation, blur, sharpen, and edge detection. Channels and alpha · Image enhancement

Special formats

Generate a Deep Zoom pyramid with sharp and the vips CLI while keeping unsupported paths explicit. Deep Zoom

Read the verification boundary

“Executed successfully” only means that the code completed on the recorded toolchain. Pixel equivalence, defaults, and encoded bytes depend on the comparison mode and differences documented on the page. See the version compatibility matrix for the baseline.

Where to go next

• Find the corresponding operation name in each interface: cross-binding operation map.
• Move a recipe into a long-running service: Production.
• Diagnose an error or unexpected output: Troubleshooting.