Developer-aware Β· Review-first Β· Cross-platform

Developer disk cleanup
with project context

reclaimit is a Go CLI and terminal UI that recognizes known build artifacts and caches, groups them by project, and lets you review a dry run before deletion.

Install audited main Β· Go 1.25.12+
go install github.com/svg153/reclaimit/cmd/reclaimit@main

Analyze, review, then dry-run

A synthetic workspace shows the complete workflow without exposing a real home directory or deleting a file.

Animated terminal demo: reclaimit finds a synthetic node_modules and Python virtual environment, reviews node_modules in the TUI, then verifies it in a dry run without deleting files
The 20 MiB belongs to generated test files, not typical user savings. Reproduce the fixture and commands.

Why reclaimit?

Most disk tools answer β€œwhat is big?”. reclaimit identifies known developer artifacts and explains how they are regenerated.

πŸ”

Developer-aware

Recognizes 17 cleanup categories across JavaScript, Python, Rust, frontend builds, Bun, pip, pipx, and macOS metadata.

πŸ›‘οΈ

Guarded cleanup

Preview before delete. --yes flag required. Context nodes that never imply deleting the repository root itself.

πŸ“Š

Reports and TUI

Plain text, Markdown with diagrams, JSON for automation, and an interactive terminal tree.

πŸ“

Repository grouping

Groups cleanup candidates by project, not just by raw size β€” so you see "next.js project caches" not just "2.3GB".

πŸ€–

Automation-ready

Export stable JSON scan and cleanup metrics for scripts, CI jobs, and reproducible comparisons.

⚑

Single binary

Prebuilt releases are standalone executables. Building current source requires Go 1.25.12 or newer.

Usage

Three commands. Three ways to reclaim space.

# Generate a Markdown report reclaimit analyze --root "$HOME" --format markdown --out cleanup-report.md
# Interactive cleanup with TUI reclaimit tui --root "$HOME"
# Clean reviewed targets (requires --yes) reclaimit clean --root "$HOME" --include-category python-venv --yes

reclaimit vs ncdu, gdu, dust, and dua

Choose by job: map the whole disk with a general analyzer, then review known developer artifacts with reclaimit.

Need Best fit Reason
Understand usage across an arbitrary disk or directory ncdu, gdu, dust, or dua General filesystem exploration and size analysis
Find known regenerable developer artifacts reclaimit 17 named categories with regeneration context
Review cleanup candidates by Git project reclaimit Repository grouping, TUI selection, and dry run
Export candidate metadata for automation reclaimit Plain text, Markdown, and JSON output

FAQ

Common questions about reclaimit.

Is reclaimit safe to use?

analyze and tui are read-only. Cleanup requires --yes, revalidates each candidate, and atomically moves it into a private same-filesystem quarantine before deletion. Changed data is preserved with a reported recovery path. Deletion remains irreversible, so review a --dry-run first and keep backups for valuable data.

What directories does reclaimit recognize?

Seventeen categories including node_modules, Python virtual environments and caches, dist, build, Rust target, .next, .nuxt, pip and pipx data, .bun/install/cache, and macOS metadata.

Is reclaimit an alternative to ncdu?

It covers a different step. ncdu, gdu, dust, and dua are better for exploring arbitrary disk usage. reclaimit focuses on recognized developer artifacts, project grouping, and review-before-cleanup. They work well together.

Can I exclude specific paths?

Yes. Use --exclude-path for exact paths or --exclude-group for prefix-based exclusions.

Does it work on Windows?

Yes. reclaimit is written in Go and supports Linux, macOS, and Windows.

How fast is it?

The scanner enforces one configurable concurrency ceiling across the complete traversal, supports clean cancellation, and reports scanned, skipped, and truncated entries. Performance depends on the filesystem, tree shape, and storage hardware.

Can I export the report?

Yes. analyze --format markdown --out report.md produces a Markdown report with tables, Mermaid diagrams, and PlantUML blocks.

Review developer disk cleanup candidates

Start with read-only analysis, inspect the result, then decide what to clean.