Frequently asked questions.
If yours is not here, ask it when you request early access.
The product
What does DataPrism do today?
We start with production Postgres databases. Your database is copied into a separate account DataPrism cannot reach, encrypted with a key from your KMS and locked for a set lock period. The copy is restored on a regular schedule, and a signed attestation records the result. Destructive changes are designed to wait until a verified copy covers them.
The product is in early access.
Does it replace my backups?
No. Keep your platform's backups and point-in-time recovery for everyday restores. DataPrism adds a copy that does not sit under the same account as production, so that one command cannot remove both.
Why not build this ourselves?
You can script a copy into locked storage yourself. The work is in what surrounds it, and that is what DataPrism is designed to take on: continuous capture that protects production, a checked and signed restore on a regular schedule, destructive changes that wait until a verified copy covers them, and keeping all of it working as Supabase, Neon and AWS change.
Which platforms do you start with?
Supabase (Postgres 17), Neon, Amazon RDS and Aurora, all three from the start of early access. The copy is stored in a separate cloud account that DataPrism cannot reach. Each promise covers only the platforms you use.
Does it watch what my agents do?
No. It is designed to control what they may destroy. Agents, scripts and CI keep the access they need, and the right to destroy moves to an executor installed next to the copy. We make no promise about what an agent reads.
Installation and costs
What does installation involve?
A copy account and its storage, keys in your KMS, organization policies and a replication role on your database, all deployed in your cloud with your own rights, from templates we hand over. During early access, we do it with you. How it works describes each step.
Can it affect production?
Continuous capture is designed to protect production first: if it falls behind, the copy records a gap and your production disk does not fill up. How it works lists the safeguards.
Who pays for storage and compute?
You do, directly to your cloud provider, for the copy and the scheduled restores. We show you an estimate of the data transfer cost before the first full copy.
How much does DataPrism cost?
Pricing is on request during early access. We bill an annual subscription per protected production environment, never per operation, per agent, per seat or per gigabyte.
Trust
Can DataPrism read my data?
We hold no key to your copy and no role in your accounts, and support never takes a dump of your database. With the default setup, the copy, the key and the compute that decrypts all sit at one cloud provider, and we never claim that this provider cannot read it. The security page lists every limit.
What happens if DataPrism stops operating?
Your copies stay locked until their lock period ends, and you can still restore them without us. The console and new releases stop. The security page explains why.
Can my auditor check the evidence without you?
Yes. The attestation is signed with a key from your KMS, and a third party can check it, and replay the restore, without a DataPrism account.
Early access
Who is early access for?
Software teams where AI agents, scripts or CI can change the production database, and where an auditor or a customer asks for proof that you can restore (SOC 2, ISO 27001, financial clients).
How do I join?
Request early access and tell us where your production database runs today. We reply by email.
Early access is open.
Bring the questions this page leaves open, and tell us where your production database runs today. Pricing on request.