DreamCleanr Clean while you sleep

AI workflow hygiene for macOS

Your Mac Wasn't Built for AI. DreamCleanr Keeps It Usable.

Reclaim storage, trim stale Docker, Claude, and Codex helper noise, and wake up to a quieter machine with receipts, protected-state rules, and balanced-safe nightly cleanup.

Download Free See What It Cleans
  • Docker-aware cleanup with protected VM boundaries
  • Claude and Codex state stays conservative by default
  • Every run leaves behind a one-page receipt you can inspect
04:30 EST Balanced cleanup scheduled Latest report ready before sunrise
Sample reclaim 5.0GB
Protected Claude VM
Trimmed 3 stale probes
GitHub-first Release-based install and update paths with public verification
Trust-first Preview-first cleanup with protected Docker, Claude, and Codex state
Free today CLI, MCP, receipts, and scheduling ship now with no fake checkout

Problem

Your AI stack is quietly eating your machine.

DreamCleanr is built for the kind of Mac that runs Docker, local AI tools, fast-moving repos, and too many terminal sessions at once.

Docker sprawl

Images, exited containers, and stale probe chains accumulate while the engine still needs protection.

Hidden cache growth

`uv`, `trunk`, Gradle, npm, and system caches keep growing long after you stop thinking about them.

Stale helper noise

Old CLI probes and forgotten helpers hang around long after the useful work is over.

Safe delete is unclear

Generic cleaners don’t understand active developer state, so you are left guessing what will break.

Solution

Smart cleanup for real dev workflows.

What ships today

Stable and trustworthy now

  • Docker-aware cleanup and safe prune planning
  • Claude and Codex process protection
  • Safe cache cleanup for regenerable developer noise
  • JSON and HTML cleanup receipts
  • Local MCP server for Claude, Codex, and VS Code
  • Balanced-safe nightly scheduling via launchd
What comes next

The next moat is deeper workflow awareness.

  • Python environment awareness
  • Node and workspace intelligence
  • Hugging Face, Ollama, and LM Studio visibility
  • Project-aware cleanup safety
  • Artifact-family disk heatmaps
  • Premium macOS shell and richer history

Onboarding Demo

See the free-to-premium flow without fake browser cleanup.

This is a sample conversion demo for the future product funnel. It shows the shape of a first run while keeping current DreamCleanr truth intact: no files are scanned in your browser, and future detectors stay clearly marked as upcoming.

Step 1 · First open

Let's see what your AI stack is hiding...

Takes about 30 seconds in this sample flow. No browser files are scanned or deleted.

Skip to free install
Step 2 · Live demo scan

Watching reclaimable surfaces pile up is the hook.

Loading demo scan...

0.0GB sample machine footprint surfaced
    Step 3 · Results

    You could safely reclaim space from known-safe surfaces.

    Read the sample receipt
    Step 4 · First win

    Start with a real local scan, not a fake browser cleanup.

    Sample first reclaim 5.0GB
    Protected state Claude VM + Docker.raw
    Next real action Install the free CLI or MCP tool
    Run your first local scan
    Step 5 · Upgrade moment

    Want guided cleanup, richer history, and premium workflows?

    • Keep the free CLI, MCP, receipts, and scheduling today
    • Raise your hand for the premium macOS shell early-access list
    • Review the grounded pricing plan before anything is sold
    Join Pro Early Access See the Pricing Plan

    How it works

    Scan. Analyze. Optimize.

    1

    Scan

    DreamCleanr inspects processes, caches, and known AI/dev storage surfaces without assuming everything is safe to delete.

    2

    Analyze

    It separates protected state, manual-review items, and cleanup candidates so you can see the tradeoff clearly.

    3

    Optimize

    Run a manual cleanup, install a balanced nightly schedule, or use MCP for preview-first cleanup from your AI tools.

    Proof

    Trust comes from visible outcomes.

    The sample report included with DreamCleanr shows a `5.0GB` reclaim, `3` stale processes trimmed, and protected Docker and Claude state left intact.

    Sample receipt balanced • apply
    Storage reclaimed 5.0GB
    Processes trimmed 3
    Objects pruned 4
    Protected Claude VM
    Manual review Docker.raw

    Why it is safer than a generic cleaner

    DreamCleanr shows what was removed, what was protected, and what still needs human review instead of hiding those decisions.

    Public verification GitHub Releases + install smoke + ops health
    Install path Evergreen, release-based, and GitHub-first
    State protection No auto-delete for Claude VM or Docker raw storage

    Install

    Choose your launch surface.

    CLI for macOS

    The fastest path for direct cleanup, receipts, and scheduling on your local machine.

    curl -fsSL https://raw.githubusercontent.com/jlsport18/DreamCleanr/main/scripts/install.sh | bash

    MCP tool for Claude

    Add DreamCleanr as a local tool so Claude can scan and preview cleanup safely.

    {
  "mcpServers": {
    "dreamcleanr": {
      "command": "python3",
      "args": ["-m", "dreamcleanr.mcp_server"]
    }
  }
}

    Tool for Codex

    Connect DreamCleanr to Codex through MCP and keep cleanup preview-first.

    [mcp_servers.dreamcleanr]
command = "python3"
args = ["-m", "dreamcleanr.mcp_server"]

    VS Code

    Use DreamCleanr as a local MCP server from VS Code’s MCP-aware chat surfaces.

    {
  "servers": {
    "dreamcleanr": {
      "command": "python3",
      "args": ["-m", "dreamcleanr.mcp_server"]
    }
  }
}

    Update DreamCleanr

    Keep the latest stable release installed without pinning yourself to a specific tag.

    curl -fsSL https://raw.githubusercontent.com/jlsport18/DreamCleanr/main/scripts/update.sh | bash

    Latest release

    Prefer release notes and direct artifacts? Jump to the current stable GitHub release.

    https://github.com/jlsport18/DreamCleanr/releases/latest
    MCP setup

    Need step-by-step client setup?

    Use the dedicated MCP setup guide for Claude, Codex, and VS Code instead of hunting through scattered examples.

    Open the MCP setup guide
    Search-ready support

    Start with the FAQ and grounded comparison pages.

    DreamCleanr now has dedicated FAQ and comparison surfaces for common search-intent questions and product-fit checks.

    Read the FAQ See the comparison

    Safety

    Built to keep the risky stuff out of default cleanup.

    Protected by default

    DreamCleanr protects `~/.codex`, `~/.claude`, the Claude VM bundle, and Docker raw VM storage by default.

    Preview-first

    The MCP surface is intentionally preview-only for destructive work, and the CLI keeps receipts first-class.

    Scheduled, not reckless

    Nightly automation stays on the balanced-safe path instead of chasing aggressive reclaim at any cost.

    Grounded roadmap

    Future Python, Node, Hugging Face, Ollama, and LM Studio intelligence is planned next, not hidden inside fake “AI cleanup” claims.

    Pricing

    Free core now. Premium convenience next.

    DreamCleanr keeps the CLI, MCP, receipts, and scheduling free. The paid layer is the later premium macOS shell, not artificial limits on basic trust-first cleanup.

    See the Pricing Plan

    Final CTA

    Stop guessing what to delete.

    Start with the free release today, see exactly what DreamCleanr cleans, and follow the premium macOS shell as the next moat.

    Download DreamCleanr Free See the Pricing Plan