confidential.storage -- Storage that proves who used your data

The thesis

Three primitives. One audit-grade receipt for every access.

confidential.storage sits in front of S3-compatible storage. Every read or write produces a portable artifact your compliance team, your customer, and their auditor can each verify on their own.

Content-addressed storage

Every object is referenced by its cryptographic hash. Your data has one canonical name no vendor can quietly rewrite -- including us.

Attested-compute gating

Per-object decryption keys release only on a verified attestation that matches the policy you wrote. The receipt schema is TEE-family-agnostic by design -- tell us what hardware you run and we co-design the integration with you.

Portable signed receipts

Each access emits an Ed25519-signed JSON receipt with the CID, attestation, policy, actor, and timestamp. Re-verifiable without us.

What we are not

A short list of things confidential.storage does not claim to be.

Clarity matters more than category-stretching. Confidential compute restricts access; confidential.storage proves the access happened correctly. The two are complements, not substitutes.

NotA confidential-compute platform. We do not run your model or your workload.
YesA storage layer that issues receipts every time confidential compute touches your data.
NotA SIEM or log-aggregation product. Receipts are artifacts, not a feed.
YesA portable evidence format your auditor verifies independently of any vendor, including us.
NotA replacement for your KMS or your IAM stack.
YesA policy-aware key-release flow that integrates with AWS KMS, HashiCorp Vault, and customer-managed keys.

Why teams pick Akave

Verifiable by design. In production.

The storage layer powering confidential.storage already runs under operators in healthcare, edge AI, and verifiable data marketplaces.

Our mission has always been to make healthcare data both accessible and accountable. With Akave, we finally have a storage layer that matches that vision -- verifiable by design, s3-compatible.

Boka

CTO, Asterisk Women's Health

Performant edge AI systems need fast, secure and verifiable file storage. Our work with Akave expands the network's ability to deliver data at dramatically lower costs.

Trevor Branon

CEO, 375ai

FAQ

Questions a CISO and a controller usually ask.

If yours is not here, the feedback form has a free-text field and we read each one.

Where do receipts live, and who can read them?

Receipts are written back to Akave, content-addressed, and signed. You own the signing key or you bring your own KMS. Your auditor can read each receipt with a public verifier and a public key; no Akave account required.

Which TEE families are supported?

TEE-family integrations are part of the design-partner co-build -- we're not announcing a fixed ship order. The receipt schema is attestation-family-agnostic: same CID, same signature, same offline verifier whether the runtime is SEV-SNP, TDX, Nitro, or something else. Tell us what you run on the form below and we'll prioritize accordingly.

If we leave Akave, do these receipts still verify?

Yes. The receipt schema is published as an open specification and the verifier tool is open-source. The CID and the Ed25519 signature do not depend on us being the host.

How does this fit alongside our existing KMS?

confidential.storage holds per-object keys in your KMS or HSM. Key release is gated on attestation + policy match. We do not custody plaintext keys for you.

Is this S3-compatible?

Yes. Reads and writes use the S3 API. Receipts are emitted alongside the response and are addressable by CID for retrieval and downstream evidence handoff.

Why are you collecting feedback before pricing it?

We have a working prototype and a clear thesis. What we do not yet have is your audit calendar, your evidence pack, and the line items your controller actually wants replaced. That is what this conversation is for.

Background reading