What SendShelf Deliberately Doesn't Do
Most tools describe what they do. This post is about the things SendShelf refuses to do, because in this particular corner of the world the refusals are more informative than the features.
We make no automated requests to Amazon. At all.
Not scraping. Not “light” scraping. Not one request to fetch a product title from a listing you pasted, which is the version that sounds harmless and is the one worth being clearest about.
When you add a product, you paste an ASIN or a product URL. SendShelf pulls the ASIN out of that string with a pattern match — a regular expression, running on our own server, against text you supplied — and then stops. The title, the brand, the category, the price: you type them. It takes about twenty seconds.
We do not fetch the listing to fill them in for you. We could. It would be a better demo. We decided against it and then went further than deciding.
The decision is enforced by a test that fails our build
A rule that lives in someone’s head is a rule until the week someone is busy. So this one lives in the test suite instead.
Our backend has an automated check that scans every source file for any HTTP client — the standard Java ones, the Spring ones, the common third-party ones — and fails the build if it finds one. Not a URL check, which could be defeated by assembling an address at runtime, but a check for the capability. The reasoning is that a URL can be constructed from pieces at runtime and an import cannot: if there is no HTTP client anywhere in the backend, there is no outbound request to any host, and that includes the one we care about.
The practical consequence is that adding a feature which called Amazon would not be a code review someone might wave through. It would be a red build, on a check whose failure message says exactly why it exists. Somebody would have to delete that test on purpose, which is a much harder thing to do quietly than adding a line to a service class.
We never ask for your Amazon credentials
There is no field anywhere in SendShelf for an Amazon password, no “connect your account” button, no OAuth flow, no session cookie import, no browser extension. There is nowhere to type one, which is the strongest form this promise can take: we cannot mishandle a credential we have no way to receive.
Your Creator Hub is not read. Your earnings are not imported. Your published video list is not synced. Everything the application knows about your products, you typed in.
Why we are this rigid about it
An influencer account is somebody’s income. Losing it is not an inconvenience to route around — it is the end of the business, and it is not obviously recoverable.
Automated access to an Amazon account is against Amazon’s terms. Any tool that offers to log in on your behalf, scrape your dashboards, or “sync” your videos is asking you to accept that risk in exchange for saving typing, and it is asking you to accept it on an account whose value is far greater than the time saved. The asymmetry is not close. Twenty seconds per product against the whole operation is not a trade any sensible person makes, so we removed our own ability to offer it.
This is also why the tool is shaped the way it is. Because we are never going to fetch anything, the product had to be worth using on manually entered data alone — which forced us to be sure the actual value was in the countdown and the ordering rather than in convenience. We think that is the right thing to have been forced into.
We are not affiliated with Amazon
SendShelf is an independent tool built for creators in the Amazon Influencer Program. It is not affiliated with, endorsed by, sponsored by, or connected to Amazon in any way. Amazon and all related marks are trademarks of Amazon.com, Inc. or its affiliates.
We say that in the footer of every page, and we mean it as information rather than as a disclaimer someone made us write. There is no partnership, no integration, no approval, and no special access. We are people who built a tool for a job, and describing who that job belongs to is the only claim we are entitled to make.
We are not a discovery platform
SendShelf will not tell you what to film. There is no product database, no trending list, no category research, no brand marketplace, and no campaign matching. It has no opinion about which products are worth reviewing and no data with which to form one.
It starts once you have decided. You tell it what you bought; it makes sure you film it before the money stops being refundable.
We are not a video editor, host, or uploader
It does not touch your footage. No editing, no storage, no rendering, no scheduled publishing, no upload automation. It records that editing happened and moves the card; the editing happens wherever it already happens for you.
The one place it touches your published video is asking for the URL and the date, and that is on purpose: a video only counts as live once both exist, which is what makes the list of everything you have published complete without anybody maintaining it.
We do not report earnings
SendShelf does not know what a video made. It cannot import an earnings file, it does not join anything on ASIN, and it draws no charts.
Every state change is recorded with a timestamp from the first day, so the data that would answer “how long does this actually take me” is accumulating whether or not we ever build the screen. But it is not a feature today, and a roadmap item is not a feature. If earnings analysis is what you need right now, you need something else right now.
What we do instead
The list of refusals is long, and what is left is small on purpose: a board that counts down every return window in your own timezone, a queue that decides what to film next and will not let you reorder it, a one-at-a-time production limit, and a cost ledger that fills itself in as you work.
That is the whole tool. If it sounds like it might be the tool for you, we were just as specific about who that is — and if it is, create an account. SendShelf is invite-only for now, and every account is reviewed by hand before it is activated.