Verified era1 archives, a permanent blob archive, and drop-in L2 node snapshots — generated, checksummed and served from hardware we own, in a datacenter we run. Direct uplink, resumable downloads, no CDN middleman.
Browse the archive Quick startEthereum clients are pruning pre-merge history (EIP-4444), and blob data expires network-wide after ~18 days. Independent, verifiable archives are how the next generation of nodes bootstraps — and how rollups, indexers and researchers look backwards.
The complete proof-of-work chain, genesis → merge block 15,537,393, in the standard e2store/era1 format every major client imports. Cross-checked against independent mirrors and re-verified file by file before publishing.
Post-Fusaka, blobs vanish from the network after the retention window. We archive blob sidecars continuously and keep them queryable long after the p2p network forgets — for rollup fault proofs, indexers and forensics.
Consensus-layer archives in era format, self-verifying via the beacon state's historical summaries. Full CL history from genesis, exported from our own archive node.
Mirror of the official Base reth snapshots — pruned and archive — with sha256
sums and a stable latest pointer.
A second, independent download source for the Base operator community.
Nitro database snapshots mirrored on the same terms: verified, resumable, and served at full line rate.
Daily Forest-format mainnet and calibration snapshots are served by our sister operation, Stax Infra — same datacenter, same operating principles.
Everything we publish is independently verifiable — that's the point of the formats we serve.
era1 files embed an SSZ accumulator checked against the canonical pre-merge header accumulator; era files verify against the beacon state. Corruption and tampering don't survive an import.
Our sha256 manifests are diffed against independent community mirrors before anything goes public, and every file is re-hashed on our side after ingest.
We permanently seed the official era1 torrent from the same box that serves HTTPS, so the community distribution stays healthy even if every web mirror disappears.
BlockX Labs runs on hardware we own in a purpose-built Norwegian facility — powered by hydroelectricity, cooled by Nordic air, and connected straight to the wire.
52-core storage server with a 79 TiB ZFS pool (dual-parity, monthly scrubs, end-to-end checksums on disk as well as on the wire).
10 Gbit dedicated to data serving, 2×25 GbE internally. Multi-hundred-gigabyte archives don't cache well on free CDN tiers, so we don't pretend they do.
Run by engineers who operate production blockchain infrastructure daily. We read the source of what we serve, and we run the same software on our own nodes.
Questions, mirror requests, or want a dataset we don't serve yet? We answer.
data.blockxlabs.ai → ops@blockxlabs.ai