Skip to content

How it runs.

Dianthan is operated as a single managed service: requests served at the edge, data stored redundantly and backed up continuously, and every product isolated from every other. This page describes how, and what we commit to.

Delivery.

Latency should depend on where your users are, not on where your account was created.

  • Served at the edge

    API requests are handled at the location nearest the caller, across a global network of data centres.

  • Encrypted in transit

    Every connection uses TLS. Plain HTTP is redirected, and sign-in links and tokens never travel unencrypted.

  • Limits per product

    Each product has its own request quota, and sign-in endpoints have stricter limits per visitor, so one busy product cannot slow another.

Data.

Stored redundantly, recoverable to a point in time, and always yours to take elsewhere.

  • Redundant storage

    Records are kept in a managed relational database whose storage is written to several locations before a change is confirmed. Files are kept in redundant object storage.

  • Continuous backups

    Changes are backed up continuously and can be restored to a point in time. The recovery window for your account is stated in the dashboard.

  • Encrypted at rest

    Data and backups are encrypted at rest. API keys and refresh tokens are stored only as hashes.

  • Export at any time

    Every product's users, records, files and logs can be exported in documented formats, without asking us.

Isolation.

Many products share the platform. None of them can see each other.

  • Scoped at the data layer

    Every row carries the product it belongs to, and every query is filtered by it. A credential for one product is refused by every other.

  • Tested on every release

    A dedicated test suite attempts cross-product access through every route before any release ships.

  • Least privilege

    Keys are scoped to a product and a set of permissions, can expire, and can be revoked at once. Customers of a product can reach only their own data.

Operations.

We tell you what is happening before you have to ask.

  • Public status

    Availability for each system is published at status.dianthan.com, with the history behind it.

  • Incident reports

    Every incident that affects customers gets a written report: what happened, why, and what changes as a result.

  • Audited administration

    Our own administrative actions go into the same append-only audit log as yours. Access to customer data is limited to named engineers and only for support you have asked for.