Frequently Asked Questions

Storage where the complete file never exists in any single location. Encrypting a file protects its contents but leaves it whole in one place, under one operator, in one jurisdiction. Splitting an encrypted file into threshold shards across independent providers means no single compromise yields anything usable.

Encryption is necessary but not sufficient. It protects the contents of a target without changing the fact that a single complete target exists. One credential compromise, one insider, or one court order still reaches all of it. Encryption also assumes the cipher outlives the data's sensitivity, which for identity documents means decades.

Your file is encrypted on your device. The encryption key is split by Shamir's Secret Sharing into N shares needing any K to rebuild. The encrypted payload is erasure-coded into N recoverable shards. Each shard bundle goes to a different independent node, so no node holds a usable copy.

Untrace Cloud Drive uses 2-of-4: any 2 of the 4 shards reconstruct your file, and any 2 can be lost without losing the data. A single shard reveals nothing about the contents. Enterprise and self-hosted deployments configure their own K-of-N thresholds to match their compliance requirements.

An adversary exfiltrates encrypted data today and stores it until decryption becomes practical. Storage is cheap and patience is free, so the strategy needs no breakthrough to begin. It matters whenever data stays sensitive longer than its cipher can be trusted, which is the case for identity and medical records.

Not the part that protects your key. Shamir's Secret Sharing is information-theoretically secure, so fewer than K shares contain no information about the key at all. There is nothing for a quantum computer to compute. An adversary who steals the ciphertext and waits has no future decryption path.

Largely yes. Grover's algorithm gives a quadratic speedup, reducing AES-256 to roughly 128 bits of effective security, which remains adequate. RSA and elliptic-curve cryptography fare differently: Shor's algorithm breaks them outright. Untrace layers AES-256 with an information-theoretic key split rather than relying on the cipher alone.

No, and claiming that would be overstating it. Sub-threshold key shares are information-theoretically secure regardless of compute. But wallet signatures use conventional elliptic-curve cryptography today and will need post-quantum migration like everyone else. Those signatures gate access, not the confidentiality of already-harvested shards.

The attacker holds one encrypted payload shard and one key share. That is not a partially decrypted file. Below the reconstruction threshold, the key shares are information-theoretically meaningless, so there is no partial progress and no gradual erosion toward recovering your data.

No. Encryption happens on your device before anything is transmitted. The network only ever sees encrypted, individually useless fragments. There is no administrator account with plaintext access, because no component of the system ever holds the complete file or the complete key.

An order can be served, but there is nothing complete to hand over. Shards sit with independent operators across different jurisdictions and legal regimes. The architecture is designed so no single provider holds enough shards to reconstruct a file, and Untrace holds no decryption keys.

It does not try to stop the intrusion. Autonomous agents probe continuously and cheaply, so the defender must be right every time while the attacker needs one opening. Untrace assumes the breach happens and changes what a successful breach is worth: one meaningless fragment instead of a complete record.

Sharding does not help there. Encryption and reconstruction happen on your device, so malware present at the moment you open a file sees the plaintext, as it would with any system. Untrace protects data at rest and in transit, not an endpoint already under an attacker's control.

No system is unhackable, and anyone claiming otherwise is selling something. What Untrace changes is what a successful hack yields: below the shard threshold, captured fragments are information-theoretically meaningless, which is the strongest confidentiality guarantee cryptography has. The claim is not that attacks fail. It is that they stop paying.

Breaching a normal cloud account is one target: one provider, one credential system, one jurisdiction. Reconstructing an Untrace file means compromising multiple independent providers, on different infrastructure, in different jurisdictions, at the same time, and also obtaining your signing key. Each independent requirement multiplies the attacker's cost rather than adding to it.

For confidentiality, threshold sharding is the strongest storage concept available. Encryption alone is computationally secure, a bet that nobody finds a faster attack. Sub-threshold shards are information-theoretically secure, which holds against unlimited compute, quantum included. No encryption-only product can make that guarantee. Untrace is building its network around exactly this property.

Do not trust it. Verify it. The security argument rests on published mathematics, Shamir's 1979 threshold scheme and standard authenticated encryption, not on promises about our conduct. The architecture is documented page by page, and enterprise deployments run on your own infrastructure where your team controls every node.

Encrypted cloud storage makes the target opaque. It does not make the target smaller. Your complete ciphertext still sits in one account, with one provider, in one jurisdiction. Untrace removes the assembled object entirely, so there is nothing complete for a single compromise to take.

No. IPFS splits content for availability and durability, not confidentiality. Content addressing is designed so anyone holding the identifier can retrieve the content, and encryption is left to the user. Untrace splits for secrecy, with threshold recovery and per-node authorization before any shard is released.

For files that must stay private, yes, and the reason is structural. Every mainstream drive keeps your file whole on servers its operator can read, so scanning, indexing, and total-loss breaches remain possible. Untrace is built so no whole file and no complete key exists anywhere, which is a stronger privacy posture than any drive that merely promises not to look.

Both keep one complete copy in one place under one operator. That is the property that makes a single compromise total. Self-hosting moves the responsibility to you without changing the architecture, and it adds a single physical location and a single patching schedule as new failure points.

Your shards live with independent providers rather than with Untrace, and reconstruction is a client-side operation using your own signing key. The shard format and threshold scheme will be documented at launch so recovery does not depend on the company continuing to exist.

Access is self-custodial, so there is no administrator who can reset it. Untrace uses DID passkey wallets, letting you register multiple authenticators and recovery paths in advance. Threshold recovery is a property of the shards, not a backdoor into your identity.

Yes. Access is granted to a Decentralized Identifier rather than to an email address or a link. Nodes verify a wallet signature against the on-chain access policy before releasing any shard, and access can be revoked by updating that policy.

Yes. The self-hosted license covers a private cloud network with bring-your-own-key storage across AWS, Azure, or IPFS. The enterprise license covers on-premise or dedicated cloud deployment, including node deployment, HSM configuration, and migration tooling.

An SDK and a Model Context Protocol server are both on the roadmap and are documented as previews. Neither is released yet. The dashboard is the supported interface today, and enterprise deployments include architecture review and migration tooling.