filter.nostr.wine
filter.nostr.wine is not just another relay endpoint under the same brand. It is a filter and broadcast layer for paid nostr.wine users, built around authentication, subscription access and a sharper idea of what a relay can do for signal quality.
Why this relay matters
The filter relay is useful because it makes relay specialization concrete. One relay can store and serve events, while another can help paid users broadcast through a filtered path. That is exactly the kind of infrastructure detail that disappears when a hub only lists endpoints.
A relay page is useful only when it explains the job the endpoint is doing. filter.nostr.wine is not just the string wss://filter.nostr.wine. It is an operated place where signed events may be accepted, served, filtered, rate-limited, searched or ignored. That makes it part of the user's trust path, even when the relay never takes custody of a key or a coin.
The practical question is simple: what happens if a client depends on this relay and the relay behaves differently than expected? A note may fail to publish, a reply may not be found, a profile may look empty, a wallet-connect message may miss the listener, or a user may assume censorship where the real issue is relay selection. Relay literacy starts there.
What the metadata says
The current NIP-11 response was reachable during the June 18, 2026 check. It identifies the relay as filter.nostr.wine, advertises NIP-01, NIP-02, NIP-04, NIP-09, NIP-11, NIP-40, NIP-42, NIP-50, NIP-59, NIP-65, NIP-94, and gives readers a concrete way to compare the endpoint with other relays.
The relay advertises its software as nostr.wine relay service. Its contact field is wino@nostr.wine, and the public operator key shown in metadata is 4918eb332a41...20e55975. These fields are not decoration. They give client developers and power users a place to start when a relay breaks, changes policy or becomes important enough to audit.
The advertised limits are also part of the story. Maximum result limit is 2000, maximum subscriptions are not advertised, and maximum message length is 524288. A client that ignores those boundaries may look broken even when the relay is doing exactly what it said it would do.
Policy, limits and trust
This is not a frictionless public commons. The metadata advertises access controls: authentication required is yes, payment required is yes, and restricted writes are yes. Those flags matter because they tell the user where policy enters the publishing path.
The tradeoff is gatekeeping. A filter relay can improve signal and reduce junk, but it also moves more trust into an operator's admission rules, subscription state and filtering choices.
Good relay choice is therefore not a popularity contest. The better pattern is to ask what the relay is optimized for: broad public reach, paid access, spam resistance, search, profile discovery, wallet traffic, local community use or long-term archival storage. When the job is clear, the relay becomes easier to trust for that job and easier to replace for everything else.
Where it fits in a relay set
Use it when studying paid relay workflows, NIP-42 authentication and broadcast/filter behavior. It is a specialized companion to nostr.wine, not a general first relay for a new user.
A healthy Nostr setup usually mixes roles. One or two broad public relays help with reach. A personal or paid relay can improve durability. A search or directory relay helps discovery. A specialized relay handles wallet, group, media or filter behavior. filter.nostr.wine belongs in that map as filter and broadcast relay, not as a magic relay that solves every routing problem.
That is why this page links both the live metadata and the surrounding sources. The reader should be able to open the endpoint, compare it with NIP-11, inspect the software or project behind it, and then decide whether the relay belongs in their client, their app defaults or their operational playbook.
How to evaluate it today
Start with the relay information document, then test from a real client. Check whether publishing succeeds, whether reads return expected events, whether filters behave as expected, whether the relay appears in other people's relay lists, and whether a second client can find what the first one wrote.
Then look at failure behavior. Does the relay return clear notices? Does it rate limit loudly or silently? Does authentication fail in a way a normal user can understand? Does the operator publish contact information? Does the software version suggest active maintenance? A relay that answers these questions is easier to recommend than a relay that merely connects.
Finally, never let one endpoint become the whole plan. Nostr works because users can move across clients and relays. The moment a single relay becomes invisible infrastructure for a profile, product or payment flow, it deserves the same scrutiny as any other dependency.