Get it for macOS

Find what filled your disk and get the space back, without guessing. Indexes the whole volume with no scan cap and counts hardlinks and APFS clones once, so shared copies do not inflate folder totals. Browse any folder by real subtree size, or work from a costed list of build output, idle dependencies, package caches, Xcode derived data and Docker layers, including compacting the Docker VM disk image that pruning alone never shrinks. Deletion is confined to roots you authorize, and a guard can clean up or hand the problem to an agent when space runs low. macOS and Linux, no account or external service.


Overview

View source

What you get

A capacity dashboard, a size-ranked file browser, and a cleanup engine that tells you what it will free before it frees it.

Reclaim indexes your whole volume — every file, no scan cap — then keeps the directory totals so browsing is instant instead of re-walking the disk on every click. A full scan of a 500 GB volume with 6 million files takes a few minutes (clone-id lookup is an extra syscall per file) and runs in the background.

Finding the space

The overview ranks the directories where size actually accumulates, filtering out the pass-through parents that make a naive "largest folders" list useless. The tree browser expands in place, so you keep the whole path in view while you drill into it. Every row carries a real subtree total.

Numbers come from allocated blocks with hardlinks and APFS clones counted once, so they agree with what deleting a folder would actually free.

Getting it back

Reclaim groups reclaimable space by what it is and what losing it costs:

  • Safe — package-manager download caches. Costs one slower install.
  • Rebuildable — build output, node_modules in idle projects, Xcode derived data, Docker build cache. Costs a rebuild.
  • Careful — the Trash, unused Docker volumes. Never preselected.

Docker gets particular attention. On macOS and Windows the daemon runs in a VM whose disk is one large sparse file, so pruning frees space inside the VM while your actual free space does not move. Reclaim measures that file directly and can compact it, reporting the bytes the host got back rather than an estimate.

Watching for trouble

A guard checks capacity every minute and grades it on percent-full and absolute free space together, because 90% of a 4 TB array still leaves 400 GB while 4 GB free is an emergency at any percentage.

When the disk is under pressure it can clean up on its own — bounded to the risk tier you authorize — or start a BB thread and hand the problem to an agent with the full report. Both are off until you turn them on. Every threshold crossing and automatic action is logged where you can read it.

Agents

Reclaim registers reclaim_status, reclaim_plan and reclaim_apply in every BB thread, plus a skill, so an agent hitting a full disk mid-build can diagnose and fix it without being told the plugin exists. While the disk is under pressure it also injects the current numbers into every thread; below the threshold it contributes nothing.

What it will not delete

Deletion only happens inside roots you authorize, and a built-in list refuses your home directory, ~/Documents, ~/.ssh, /System and similar even when a root would cover them. Paths are resolved before they are checked, so a symlink inside an authorized root that points outside it is refused.

The apply step never accepts a path. It takes identifiers Reclaim minted while building a plan and re-authorizes each one immediately before removal — the panel, the bb reclaim command, and an agent all pass through the same gate. Nothing inside a macOS application bundle is ever a candidate.

Requirements

macOS or Linux. No account, service, or separate install. Docker features need a local Docker install and appear as unavailable without one. APFS clones (Chrome code-sign leftovers, Finder copies) share one data stream; Reclaim counts that stream once instead of once per name.

More from Braedon Saunders

Audio PreviewPlays .m4a and other audio files in BB's file panel, including audio/mp4 that the built-in preview refuses.18BallastKnow which thread is eating your RAM before the machine starts paging. Reads vm_stat page classes rather than the free percentage macOS reports, which counts inactive pages as free and stays reassuring while the box holds gigabytes of swap — pressure is graded on headroom, paging rate, compression and the kernel's own verdict, and the banner names the signal that decided. Consumers roll up by process tree and by BB thread, matched through worktree paths, so a 2.4GB row names a thread you can steer instead of six identical node processes. A guard can throttle BB concurrency, ask the hungriest threads to release memory, stop orphaned processes, or hand the problem to an agent — every one of those is off until you turn it on. Nothing is stopped unless nothing can still be waiting on it, and the apply step takes only ids the plan minted, re-checked against the live process table. macOS and Linux, no account or external service.6Muse CodeAdds Meta's Muse Code as a bb provider over its own session protocol: streamed timeline, approvals, subagents, model picker, health, and a subscription meter measured from Muse's session logs.29Provider UsageShows how much of each provider's plan is left — every rate-limit window, percent remaining, and when it resets — plus a 7/30/90-day token chart covering every provider, from local transcripts and BB's own usage events.85

More in Utilities 12

Inspect or control the computers bb runs on.

View all