Guides

Task-oriented how-to guides for working with sleap-io. Start with Examples for copy-paste recipes, then reach for the focused guides below.

Examples: Practical recipes — loading and saving, editing skeletons and tracks, fixing video paths, exporting to other formats, rendering, and more.

Remote loading: Load .slp files and videos straight from http/https, cloud storage (s3://, gs://, az://), and Google Drive URLs, with lazy range-based streaming, optional persistent caching, and authentication.

Codecs: Convert Labels to and from NumPy arrays and pandas DataFrames using the array and table codecs.

Merging: Combine datasets from multiple sources with track, skeleton, instance, and video matching.

Rendering: Render videos and images with pose overlays, customizable colors, markers, presets, and motion trails.

Transforms: Crop, scale, rotate, and pad labels and their videos while keeping coordinates aligned.

Reference material

For the data structures these guides operate on, see the Data model; for format-specific details, see Formats; and for the sio command line, see the CLI reference.