FAQ

Is this a blockchain?

No. There is no consensus layer, no smart-contract VM, no token, no gas. The design insight is that consensus is unnecessary when every piece of state has exactly one writer — and everything in ZephCraft does (your files, your databases, your registry names). "Shared" views are built by readers aggregating across many single-writer sources.

Is there a token?

No, and none is planned. The free tier runs on reciprocity — give disk, use disk. A paid tier, if it ever exists, would settle in ordinary USDC.

How durable is my data, really?

Honest answer: content is erasure-coded at roughly 4× redundancy (k=8/n=32) and continuously repaired — losing most piece-holders is survivable, and a publisher's own copy is only released once other nodes hold a full erasure set. But this is an early network with a small node count; treat it accordingly. The live numbers at zeph.craft.ec are real, not projections.

Can deleted data really disappear?

Two honest halves:

• Public content: you can withdraw your copies and records, and unwanted content fades — but you cannot force other nodes to delete what they want to keep. That's censorship-resistance, working as intended.
• Private content: deleting drops your key capsule and lets the ciphertext fade — a strong best-effort shred. A guaranteed cryptographic shred of already-published data is not physically possible with a locally-held key (your key is derivable from your identity as long as you hold it); the real mechanism — key shares held by independent agents — is designed but not built.

What happens if the founders disappear?

Running nodes keep storing, serving, and healing indefinitely — the protocol has no seat only the founders can hold. Content routing is a DHT anyone participates in; relays and bootstrap seeds are anyone-can-run; governance is a multisig designed to hand off to independent maintainers. What would be lost is upgrade coordination, not the network.

What's not built yet?

We keep this list honest:

• Verification (independent re-execution of program results) — designed, not built.
• Sharing encrypted content with other identities (proxy re-encryption) — the crypto is in-tree, the sharing flow is design-only.
• Guaranteed crypto-shred (Tier-0, key-share agents) — design-only.
• Storage proofs (PDP) and reputation — not built; today's pollution defense is per-piece verification tags.
• Large-file segmentation — objects are coded whole today; the streaming multi-segment path is on the backlog.
• SWIM-style failure gossip — node deaths are detected by each peer directly; epidemic dissemination comes with scale hardening.
• Open-source release — the code is source-available-on-release; opening the repo is gated on the spec/docs being presentable (this site is part of that).

How do the product apps relate?

Craftec's apps (MindCraft, HandCraft, SalamCraft, FlowCraft, …) are independent products today — most on Solana. ZephCraft is the infrastructure network they can move onto as it matures; a federated multi-owner app pattern (many sovereign databases + cross-node reads) is already validated on the network.