"Is it still up?"
This one sounds like a question about a website, and the honest answer takes the question apart. A status label on this site records whether an address answered at a point in time. That is the entire claim. It is not a stock notice, it is not a service announcement, and it does not describe what a platform intends to do next week. One reading, one moment, one label with a date attached to it.
The dated timeline behind the label is short and specific. A notice on 2026-06-01 described a security and compliance upgrade and mentioned disrupted links, slower pages and some orders cancelled. Purchasing and top-up stopped on 2026-06-12 and stayed stopped through the end of the month while warehouse and parcel work continued. The gradual restart planned for 2026-07-01 did not restore ordering, and by August third-party records show the storefront offline. No permanent closure was ever announced by the operator.
So the answer splits in two and both halves are true. The operation is down in practice, and it is unresolved on paper, because nobody with authority has said the word closed. A reader who needs one clean yes or no will not get it here, and the site will not manufacture one to make an answer tidier. What sits in the record is a set of dated readings, and a reading is a different kind of object from a verdict.
What the label is genuinely good for is comparison. Two rows read on the same date can be lined up against each other, which is how the platform ledger separates names that are still accepting orders from names that are recorded as closed. That is a narrow use for a narrow tool, and it is the use the label was built for. Nothing in it predicts the next announcement.
"Can I still buy?"
This is the question with the shortest factual answer, and it is not the answer anyone wants. Purchasing stopped on 2026-06-12. The restart announced for 2026-07-01 did not bring ordering back. Around 2026-07-10 a community representative said ordering had not been restored and offered no new date. By August the storefront was offline. There is no verified channel on this site through which a new order can be placed with the operation that used to fill these sheets.
What that changes depends on which stage a row has reached, and the stages behave very differently. A row that was never purchased can no longer become a purchase through the old route. A row that was purchased and is already paid sits with the party holding it. A row whose item reached a warehouse is the one case where something can still happen, because a parcel can still be submitted even when the buying step is gone.
The constructive half of the answer is the rebuild. Thirty-one wrapped addresses in this period's sample failed at the network layer when requested as wrappers, and the same thirty-one, rebuilt into the listings they encode, answered normally. Those rebuilt addresses are ordinary marketplace listings. A platform that accepts marketplace listings can be given one today, which is why the practical route after the shutdown runs through the rebuild rather than through the old order button.
None of that restores the ability to buy from this operation, and the distinction is worth holding onto. Rebuilding a row is not placing an order. It is recovering the address that a purchase would need, so that the address is ready whenever a reader decides to use a platform that is still accepting orders. Preparation and purchase are two different steps, and only one of them is currently blocked.
"Are my links dead?"
Dead is the wrong word for most of what is sitting in these sheets, and the sample is unusually clear about why. Thirty-one wrapped addresses collected from public sheet pages were requested exactly as they appear in the file. Not one of them connected. The same thirty-one, rebuilt into the source addresses they encode, returned a normal response. The wrappers are dead. The listings behind them were reachable the whole time.
Which layer failed matters, because it decides what a reader does next. A rebuilt address is a plain listing on a marketplace, and all ninety-three records in this period's sample carried a source marker reading WD, pointing at Weidian. There were no moved records, no gone records, no shielded records and no unrecorded records in the table. Everything that could be reached, answered in the same state.
That uniformity has a limit, and the limit is worth stating plainly rather than burying. Every marker observed here pointed at one marketplace, so nothing in this period describes how a row from a different marketplace would behave. A sheet built around other sources would need its own reading before anyone could say what its links do. The honest position is narrow: these thirty-one were rebuildable, and no other marker has evidence behind it on this site.
So the answer to the question is that the links are raw material. They are not destinations any more, because the wrapper in front of each one stopped answering, but they still carry the identifier of the listing they were written for. A reader who treats a wrapped address as a thing to be opened gets an error. A reader who treats it as a thing to be decoded gets a working address back, and the legacy link resolver is the page that performs the decode.
"Do I lose my photos?"
The photo question is narrower than it sounds, and it has two halves that get tangled together. The first half is whether new photographs can be taken, and the answer is no: the ledger row for this operation records free inspection photos as not available, with purchasing unavailable since 2026-06-12. A new order is what triggers a new photo set, and no new orders are being placed. That half of the answer is closed.
The second half is what happens to a photo set that already exists. A set belongs to the order it was produced for, and it stays attached to that order rather than following the buyer to a new platform. Nothing in this record suggests a set gets transferred, migrated or reissued on a reader's behalf. What a reader holds is the copy they downloaded at the time, which is why downloading early was never a trivial housekeeping step.
The same logic covers the rest of the record a reader might be worried about. Order history stays with the account that produced it. A balance held in that account is a matter between the account holder and the operation that holds it, and no channel this site can name is settling account money today. These are unresolved questions rather than losses, and the difference matters when a reader is deciding what to do next.
What survives reliably is whatever left the account while the account still worked: downloaded photographs, saved order numbers, exported rows, screenshots with visible dates. Those are the pieces that a reader still controls. Anything that only ever lived inside the platform is now a request rather than a possession, and the open questions hub is where this site records which of those requests have an address and which do not.
What the five have in common
Line the five phrasings up and the shape underneath them appears. One asks about a service, one about a purchase, one about an address, one about a file, one about a record. Each names a different object, and each object sits at a different stage of the same chain: the platform, the listing, the order, the item, the evidence. The word a reader reaches for first tends to be the part of that chain they are standing closest to.
That is why the five questions produce five different answers instead of one. The platform layer stopped. The listing layer did not. The purchase layer stopped. The parcel layer may still be open. The evidence layer is whatever the reader already saved. Reading them as one question and expecting one answer is how a person ends up concluding that nothing works, when in fact most of the chain is still standing and one link is missing.
There is also a single discipline that all five answers depend on, and it is the habit of dating a claim. A status label without a date is an opinion. A row without a check date is a hope. This site excluded one hundred and thirteen addresses from the last sample rather than writing them down as unknown, and that choice is the same choice a reader makes when they mark a row as unchecked instead of assuming it is fine.
Five wordings, one underlying worry, and the answer that holds for all of them: what you can check, you date, and what you cannot check, you label. The answers hub collects these questions as they arrived, and the FAQ page handles the short versions. Neither one promises that the missing link comes back. Both of them keep the record straight about which links are missing and which are still carrying weight.