The File-Sharing Layer Nobody Audits Until a Supplier Contract Leaks
Cross-border operators spend real money hardening the storefront — WAF rules, chargeback tooling, three different analytics stacks — and then email a factory cost sheet, a 3PL rate card, or a .env file through a consumer cloud drive with no expiry and no audit trail. That asymmetry is the actual story behind Aktar, a native macOS menu bar app from Mert Topuz that uploads files straight to your own S3-compatible bucket and copies the shareable link to your clipboard. No account, no intermediary server. For anyone running supplier negotiations, marketplace appeals, or agency handoffs across time zones, the interesting part isn’t the app — it’s the architectural argument underneath it.
What Problem This Actually Solves (And For Whom)
The maker’s own framing is the most honest part of the launch: he kept repeating the same loop — open an S3 or R2 dashboard, upload a file, hunt for the public URL, copy it, paste it into Slack. He built Aktar so that loop collapses into seconds, and he says so directly in the launch thread.
Strip away the developer framing and the cross-border use case is obvious. A sourcing manager photographing a defect batch on a factory floor. A brand owner sending Amazon Seller Central appeal documentation — invoices, supplier authorization letters, test reports — to a consultant. An agency sending a client the raw ad creative files before a Q4 push. Every one of those flows today defaults to Google Drive, Dropbox, or WeTransfer, which means the file lives on someone else’s infrastructure, often indefinitely, often indexed by whoever has the link.
Aktar’s answer: the file goes to storage you already pay for, and the link expires. The 0.6.0 release added temporary links at upload — 1 hour, 24 hours, or 7 days — and those work even on private buckets. The 0.9.1 release added 5 and 15 minute windows explicitly “for credentials and other files that shouldn’t live on.” That last detail is the one that should make a cross-border operator sit up. How many times has your team pasted a Stripe key, a Shopify API token, or a 3PL portal login into a shared note?
Why Amazon sellers should care more than Shopify ones
Shopify operators live in a world of apps and integrations where file sharing is mostly a marketing-asset problem. Amazon sellers live in a world of compliance documentation. Account reinstatement packages, product liability certificates, MSDS sheets, trademark registration proofs — these get emailed, re-emailed, and forwarded to agencies, VAs, and consultants across jurisdictions. The blast radius of a leaked supplier cost sheet or an unreleased product photo is materially higher than a leaked lifestyle image. Expiry-by-default matters more when the recipient list is a rotating cast of freelancers you’ll never audit.
How It Differs From What You’re Probably Using Today
The obvious comparison set is WeTransfer, Dropbox, Google Drive, and the file-sharing features baked into Slack. All four route your file through their infrastructure. All four have had their share of accidental-public-link incidents. And all four want you to think of them as the destination rather than the pipe.
Aktar inverts that. It’s a client, not a destination. It works with Amazon S3, Cloudflare R2, Backblaze B2, MinIO, and other S3-compatible providers. If you’re already paying for R2 or B2 as a CDN origin for your storefront images, you’re effectively getting file-sharing infrastructure for free — the marginal cost of a 40MB supplier PDF on R2 is rounding-error territory.
The second differentiator is the privacy posture. The 0.8.0 release strips GPS location from photos before upload by default, with options to strip all metadata or keep everything per destination. Critically, the image isn’t re-compressed, and if the location can’t be removed, the upload stops rather than leaking it. That’s the right failure mode. Compare that to the average WeTransfer upload, where EXIF data travels intact unless the recipient strips it.
The “Test Connection” detail is more important than it sounds
A commenter flagged the 0.5.2 release specifically because “Test Connection now shows each step and explains why a public link fails. That’s exactly where silent 403s bite the first time you point Aktar at a new R2/S3 bucket.” Anyone who has configured a bucket policy knows this pain. A misconfigured CORS rule or a private ACL produces a link that looks fine and returns a wall of XML. For operators who aren’t cloud engineers — which is most of them — a diagnostic that explains why the link fails is worth more than a feature. It’s the difference between a five-minute setup and a two-hour Stack Overflow spiral.
What Cross-Border Sellers Can Borrow From This
Even if you never install Aktar, there are three operational patterns worth stealing.
First, treat your storage as infrastructure, not a subscription. Most DTC brands pay Dropbox or Google Workspace per seat for storage they could get for cents on R2 or B2. The migration is annoying, but the lock-in cost compounds. If your team is already on Google Workspace for email, fine — but audit whether the file-sharing tier is earning its line item.
Second, default to expiry. The 0.9.1 release’s 5 and 15 minute windows are a design statement: most shared files don’t need to live forever. Apply the same logic to your internal ops. Vendor onboarding docs, temporary portal credentials, one-off creative review files — these should not persist in a shared drive for three years because nobody remembered to clean up.
Third, metadata hygiene is a real cross-border risk. A factory photo with embedded GPS coordinates tells your supplier’s competitor exactly where the production line is. A product sample photo taken at a trade show with location data intact is a small but real intelligence leak. The 0.8.0 metadata stripping is a feature most operators won’t think to ask for until they need it.
Folder uploads and the agency workflow
The 0.9.0 release added folder uploads — drop a folder and it goes up as one ZIP with one link, or file by file with structure intact, and hidden files like .env and .git are excluded automatically. That last exclusion is a quiet safety net. If you’ve ever watched a developer zip a project directory and accidentally ship credentials, you understand why “hidden files are left out automatically” is a feature, not a footnote.
For agencies managing multiple brand clients, this maps cleanly to deliverable handoffs. One ZIP, one expiring link, no client-side folder archaeology.
Where My Judgment Says It Falls Short
Three things give me pause.
The kill switch problem is real and unresolved. Gal Dayan raised it directly: “if I share a temp link and then realize I sent it to the wrong person, is there a way to kill it early from the app, or am I stuck waiting out the 5⁄15 min window?” The maker acknowledged it and put a manual kill switch on the roadmap. Until it ships, a 7-day link sent to the wrong email address is a 7-day liability. For the short windows this is tolerable. For the long ones it isn’t, and cross-border operators routinely need the long ones.
The platform story is thinner than it looks. The launch thread says “Windows and iOS are already out too,” and the maker has stated plans for Android next. But folder uploads and metadata removal are explicitly “coming to Windows next” per the changelog. If your ops team is Windows-primary — and in a lot of 3PL and warehouse environments, it is — you’re getting a materially less capable product today.
No team accounts, no audit log. This is the gap that would stop me from recommending it for anything compliance-adjacent. If you need to prove who shared what with whom and when — which you do the moment a supplier dispute escalates or a marketplace asks for documentation provenance — Aktar doesn’t give you that. It’s a personal utility, not an enterprise control plane. That’s a fair design choice, but it caps the addressable use case at “individual operator sharing a file” rather than “team managing a document lifecycle.”
Where the math breaks
The pitch is “no intermediary server,” which is true and good. But the corollary is that you are now the sysadmin. Bucket policies, CORS, lifecycle rules, access keys — that’s your problem. For a solo operator with an existing R2 account, fine. For a five-person brand team with no engineer, the setup cost may exceed the Dropbox subscription they’re trying to escape. The 0.5.2 Test Connection improvements narrow that gap, but they don’t close it.
What I’d Watch / Test Next
Three concrete things to do this week, depending on where you sit.
If you already run R2 or B2 for storefront assets: install the Aktar CLI via Homebrew (brew install getaktar/tap/aktar-cli) or the npm package, point it at a scratch bucket, and run one real supplier-document handoff through it. Measure the setup time honestly. If it’s under ten minutes, you’ve found a replacement for at least one paid tool.
If you’re on Windows or Android: wait. The feature parity isn’t there yet, and you’ll be frustrated.
If you’re a team: don’t adopt this as infrastructure. Adopt the principle — expiry by default, metadata stripped, storage you control — and apply it to whatever tooling you already have. The full changelog is worth reading as a case study in how a solo maker ships fast against user requests; four releases in a single launch day is a pace most SaaS roadmaps can’t match. Watch for the kill switch and team features. Those two additions would move this from “clever personal utility” to “defensible ops layer.” Until then, it’s a sharp tool for a specific operator profile — and a useful reminder that the file-sharing habits of most cross-border teams are years behind their storefront security.
