s3m

A command‑line utility enables direct storage of data streams into S3‑compatible buckets. It accepts input from files, pipes, or standard input and performs multipart uploads that can be resumed after network interruptions, avoiding the need to restart large transfers such as database backups. The tool operates with minimal resource consumption, requiring only a few hundred megabytes of disk space for buffering and avoiding full in‑memory loading of the stream.

It supports configuration for multiple S3 providers, allowing users to define endpoints, regions, and credentials in a simple YAML file. Users can upload, download, list, and delete objects, as well as create and remove buckets, with optional JSON output for scripting. Additional commands provide usage summaries, grouping by time periods, and management of resumable upload state.

The design targets constrained environments where low memory and disk usage are critical, while offering fault‑tolerant, resumable uploads for reliable handling of large or continuous data streams.