AutoSort

Modern users often accumulate hundreds of unorganized files in folders like Downloads, leading to clutter, lost documents, and wasted time. Manually organizing files is repetitive, error-prone, and rarely maintained consistently. AutoSort is a privacy-first autonomous agent built on OpenClaw that tackles this problem through intelligent local execution. The agent scans a chosen directory, examines file names, extensions, and metadata, and uses multi-step reasoning to decide how to categorize and rename each file. It then performs real file system actions, creating folders, moving files, resolving conflicts, and detecting duplicates. What sets AutoSort apart from traditional scripts is its persistent memory and adaptive learning. The agent remembers user preferences across sessions for example, learning that career-related PDFs belong in a specific folder and uses this knowledge to make better decisions in the future. It also includes self-correction logic i.e if a move fails or a conflict occurs, the agent retries using alternative strategies. Additionally, AutoSort supports a long-running watch mode, allowing it to monitor folders and proactively maintain organization over time. All operations are executed locally through OpenClaw, ensuring full privacy, user control, and reproducibility. This project demonstrates meaningful autonomy, practical tool usage, persistent memory, and real-world utility, showing how OpenClaw agents can deliver tangible everyday value.