Primary Mirror
Activehttp://nexusbj3sjjjorrg2jnbdlf5xktaapk7rawiovmdpoimf5aiet4a2zid.onion
NexusVendorPortal
Mirrors operational
NexusVendorPortal
Mirrors operational
Three signed v3 onion addresses. Each one resolves to the same Nexus Market backend through encrypted tunnels and shares the same PGP key for signature verification. Pull from Copy, never from memory.
The three onion addresses above are operated independently from the market core. They sit behind separate hosting providers, are reachable through separate Tor circuits, and run their own independent rate limits. A coordinated takedown of one mirror has no effect on the other two. Latency drift between them is normal and reflects which exit nodes Tor is currently routing through, not load on the market itself.
Once every 24 hours the gateway custodian signs a fresh timestamp block with the master key, fingerprint 0A9D. The block is published on each mirror login page and on this gateway. Importing the public key once and verifying each new login is the difference between a clean session and a credential donation.
Mirrors are retired only when latency rises above two seconds for a sustained 30-day window or when a hosting provider issues a takedown notice. New mirrors are published here first, signed with the same key, and given a 72-hour overlap with the retired address before it is dropped from the roster.
Run this on a fresh checkout of the master public key. The key file is published on the About page and signed by independent witnesses, you do not need to trust this gateway alone.
If GPG complains about untrusted certificate path, that is expected on first run. The signature is still cryptographically valid, the warning is about trust web depth. After verifying out of band that the key fingerprint matches, you can sign it locally to silence the warning.
The probability of all three v3 onions failing simultaneously is low but not zero. When it happens, the cause is usually upstream. Tor itself runs into directory consensus issues a few times a year, exit nodes get rerouted, hidden service descriptors fail to publish for a few minutes. None of this means the market is gone.
The right move is to wait. Five to ten minutes of patience clears 90 percent of these incidents. If the outage persists past 30 minutes, reload this gateway. The roster updates faster than the market core does, and a new fallback address will appear here if the rest of the set has been compromised. Do not search for replacement URLs on the clearnet, that is exactly when phishing networks pump fake mirrors into search results.