The Real Cost of “Clean” on a Seller’s Machine
Every cross-border operator I know runs their business on a laptop that is quietly suffocating. Between Amazon bulk upload sheets, Helium 10 exports, Figma mockups for a Shopify theme refresh, a Docker container for the feed-sync script, and a node_modules folder for the TikTok Shop API integration, the disk fills up long before the business does. The tools that promise to fix this have always felt wrong for our use case: they want one giant “Clean” button, they don’t understand developer artifacts, and they charge a subscription for the privilege of guessing. So when I saw Shedi launch from solo maker Otabek Abdiraimov, I didn’t file it under “consumer Mac utility.” I filed it under “the unglamorous infrastructure that keeps a one-person brand from stalling on a Tuesday afternoon.” Here’s why a disk cleaner matters to a seller, and where I think this one is still thin.
The Problem Shedi Actually Solves
The pitch is narrow and honest: the maker’s MacBook kept running out of space, and the culprits were always the same — Xcode simulators, DerivedData, Docker images, and node_modules. The big-name cleaners didn’t know what any of that was, cost more, and asked him to trust a single magic button. On his own machine, Shedi’s first scan found 21.5 GB. That number is the whole thesis. For a developer, 21.5 GB is a nuisance. For a cross-border seller running a two-person DTC brand, 21.5 GB is the difference between being able to install the next macOS update and not.
What Shedi does instead of a magic button is enumerate. It claims awareness of 20+ developer tools’ hiding spots — Xcode, Docker, node_modules, npm, pnpm, Homebrew, Cargo, Gradle and more — and for every item it tells you what it is, whether it’s safe, and whether it comes back. Only artifacts that rebuild themselves are pre-selected; your own files never are. Everything routes to the Trash first so it’s recoverable. No account, and files never leave the machine. There’s a playable demo on shedi.app before you download anything, and an early price where the first 50 buyers get $10 off at checkout, with the site showing how many slots remain.
Why this is a seller problem, not just a developer problem
Most Amazon FBA brand owners I talk to don’t think of themselves as technical. But look at what’s actually on their machine by year two: a Shopify CLI install for theme work, a Python environment for the inventory-forecasting spreadsheet someone built, a Docker image for the repricing bot, and node_modules from the one time they tried to wire Klaviyo into a custom landing page. None of that shows up in a consumer cleaner’s vocabulary. The result is the classic pattern — the operator buys a bigger drive, or offloads to iCloud, and the underlying waste never gets addressed. A tool that names the waste is worth more than a tool that hides it.
How It Differs From the Incumbents
I’ve used the obvious comparison points. CleanMyMac X is the category default and it’s fine at what it does — system junk, mail attachments, large-file hunting — but its developer-tool coverage has always been shallow, and it’s priced as a subscription for a job you do a few times a year. DaisyDisk is a beautiful visualization tool and I still recommend it for finding the one giant folder you forgot about, but it shows you where the space is, not what is safe to delete. OnyX is free and powerful and utterly opaque to anyone who isn’t already a Mac admin. And the “just run a script” crowd will tell you to rm -rf node_modules and move on — which works until you delete the wrong project’s dependencies and spend an afternoon reinstalling.
Shedi’s differentiation is the explanation layer. Every item says what it is and whether it comes back. That’s a small UX decision with an outsized effect on trust, because the failure mode of every cleaner is the same: you don’t know what it removed until something breaks. Routing to Trash first is table stakes but table stakes that a surprising number of competitors skip.
Why Amazon sellers should care more than Shopify ones
Shopify operators tend to work in the browser and a theme editor; their disk pressure is mostly images and video assets, which any cleaner handles. Amazon sellers are the ones running bulk flat-file operations, Keepa exports, ad-report CSVs, and increasingly a local script or two to reconcile inventory across Seller Central and a 3PL. That’s the profile that accumulates node_modules and Docker layers without ever intending to. If you’re an FBA brand owner who has ever muttered “why is my Mac full” while staring at a 512 GB drive, this category is aimed at you whether or not you write code.
What Cross-Border Operators Should Borrow From This Launch
Three things, and none of them are about disk space.
First, the honesty about what the tool doesn’t do yet. Shedi isn’t notarized, and the maker says so up front, with a 30-second Open Anyway guide and a promise that the notarized build is a free update. That’s how you ship a v1 to a skeptical audience. Cross-border sellers launching a new SKU or a new storefront could learn from it: name the limitation before the customer finds it. The refund request you prevent is worth more than the feature you ship.
Second, the “explain every item” pattern. The reason Shedi’s approach works is that it converts an opaque action into a reviewable list. That’s the same principle behind a good returns policy or a transparent fulfillment SLA — the customer (or the operator) needs to see the reasoning, not just the outcome. If you’re building any internal tooling for your brand, steal this: never automate a destructive action without a human-readable manifest.
Third, the pricing model. A one-time early price with a visible counter of remaining slots is a clean way to launch to a small audience without pretending you have a subscription business. For DTC operators testing a new market or a new channel, the equivalent is a time-boxed pilot with a stated cap — not an open-ended “we’ll see how it goes.”
Where the math breaks
The pnpm question in the launch comments is the one I’d flag hardest. A commenter, Lukáš Gurecký, asked whether Shedi accounts for pnpm’s hard-linking, since most of node_modules is linked to a global store and deleting the folders frees less space than their reported size suggests. That’s a real accounting problem, and the launch page doesn’t answer it. If Shedi reports 8 GB of node_modules but only 2 GB is actually reclaimed, the trust you built with the explanation layer evaporates in one scan. The maker also hasn’t publicly addressed the related ask — preselecting node_modules in projects untouched for 30 days while keeping active ones — which is the feature that would make this genuinely useful for anyone juggling multiple storefront codebases.
Where My Judgment Says It Falls Short
I like the philosophy and I’m skeptical of the execution maturity. Shedi is a solo project, not notarized at launch, and the developer-tool coverage is a claim rather than a verified list — I’d want to see it correctly identify a Gradle cache and a Cargo registry on a real machine before I trusted the “safe” labels. The Trash-first design is the right call, but it also means the tool doesn’t actually free space until you empty the Trash, which is a UX trap for the exact non-technical seller I described above.
The bigger strategic question is scope. Shedi is a Mac-only utility aimed at developers. The cross-border seller who needs it most is often on Windows, or on a company-issued machine they can’t install unsigned apps on. And the category it’s entering is crowded with free options — OnyX, the built-in Storage Management pane, and a dozen open-source scripts. Shedi’s bet is that explanation and safety are worth paying for. I think that bet is right for a slice of the market, but it’s a slice, not a wave.
The notarization caveat deserves more weight
The “Open Anyway” flow is fine for a developer. For a seller who’s been trained by years of phishing warnings to never click through a security prompt, it’s a conversion killer. The maker says the notarized build is a free update, which is the right answer, but until that ships, the addressable audience is people who already trust the maker — and that’s a small circle. If you’re evaluating Shedi this week, wait for notarization unless you’re comfortable with the workaround.
What I’d Watch / Test Next
This week, if you run a Mac-based operation, do the boring thing first: open System Settings › General › Storage and screenshot what it reports. Then run the Shedi demo and compare its 20+ tool categories against what you actually have installed — Xcode, Docker, a node project, a Homebrew stack. The gap between the two lists tells you whether the tool is aimed at you. If you’re a seller with a single Shopify store and no scripts, you probably don’t need this; DaisyDisk will do. If you’re an FBA operator with a local repricing script and a Docker container, the 21.5 GB claim is worth testing against your own machine.
Watch two things over the next quarter: whether the notarized build ships on the promised free-update path, and whether the maker answers the pnpm hard-link accounting question publicly. The first determines whether non-technical sellers can adopt it; the second determines whether the numbers it shows are trustworthy. A cleaner that overstates reclaimed space is worse than no cleaner at all, because it teaches you to trust a number that isn’t real — and in cross-border ops, trusting the wrong number is how you end up with a stockout.
