Skip to content
Bifröst Orbit
Sign in

Trust centre

Your data, in safe hands.

Organisations that answer for what they publish need to know where their data lives, who can reach it and what happens when something fails. Here is how Bifröst Orbit looks after it, what you can count on when something goes wrong, and how we test it.

At a glance

Everything in one place

Where your data lives

In the EU, with every provider we use listed.

Report a security problem

Email us what you found, and a person on our team replies.

Data processing agreement

Our agreement under Art. 28 GDPR, with the measures we take.

Privacy policy

What we collect, why, how long we keep it and your rights.

Accessibility

Our WCAG 2.2 AA target and how we test.

System status

How each service is doing right now, and recent incidents.

Open

Check it yourself

The code, packages and API documents anyone can read, without asking us.

Code on GitHub

Our SDKs and starter projects are public, and you can report a vulnerability on any of them.

CMS SDK on npm

@bifrost-orbit-cms/sdk, for websites and apps that read content from CMS Orbit.

Careers SDK on npm

@bifrost-orbit-careers/sdk, for career sites that show jobs from Careers Orbit.

API catalogue

Every public API with its OpenAPI document, also listed in apis.json and at /.well-known/api-catalog.

security.txt

Who to contact about a security problem, in the standard format.

Security

Security is part of the platform, not an add-on

Single sign-on

Connect Microsoft Entra ID, Google Workspace or any OpenID Connect or SAML 2.0 provider, such as Okta. Prove your email domains with a DNS record first, so nobody can claim another company's domain, then require it for everyone on them.

Passkeys, no passwords

People sign in with a passkey or an email code. Orbit keeps only the passkey's public key, and faces and fingerprints never leave the device. Signing out ends the session on our servers, so a copied session can't be used again.

Access in one place

Orbit decides who belongs to your organisation. Remove someone once and they lose access to every product, or let your identity provider do it with SCIM.

Roles and permissions

Built-in roles, custom roles on Growth and above, and access that can be limited to regions. Every permission is checked on the server.

Audit trail

Who changed what, when and on whose behalf, including AI agents. Orbit records every sign-in attempt, and every time our own team opens your organisation.

Keys that stay secret

Secret keys are rejected in browser code, production keys can't read drafts and webhook messages are signed. Only admins can remove revoked keys.

Organisations kept apart

Every record carries its organisation, and data is always read and written within it.

Safe by design

Signed-in apps only run the scripts we serve, and addresses you give us, such as webhooks, can't be used to reach our own network.

Checked on every change

Every change to our code is checked for vulnerable dependencies and for secrets that don't belong in it. Our servers accept administrative access only with keys, from known addresses.

Deletion

Mistakes you can undo

Deleting a page, file or website moves it to the bin, where it can be restored. By default, only the Owner can erase anything for good, and deleting a whole organisation has to be confirmed in Orbit before anything is erased.

AI agents

Agents work inside your rules

An agent acts for one person and never has more rights than they do. It can't publish, change protected prices or settings, or read the audit trail, and every page change it makes needs a person's approval.

Resilience

Resilience is built into the foundation

Bifröst Orbit is designed to keep services available when infrastructure or individual applications run into problems. Automatic failover, database replication, independent satellites and resilient event delivery keep the platform running when something goes wrong.

Hosted in the EU

The platform and its data run in IONOS data centres in the EU, with a recovery copy of the database and its backups with STRATO in France. Files and backups are stored in Amazon S3 in Ireland.

ISO 27001 certified hosting

Our hosting provider, IONOS, is certified to ISO/IEC 27001:2022 by TÜV NORD for operating the infrastructure in its data centres.

Automatic failover

A standby server keeps a live, replicated copy of the database and takes over by itself if the main server fails. We test the switch in production.

Backups every hour

Encrypted, versioned backups every hour and every night, kept away from the servers they protect, and restored on purpose to prove they work.

Protected at the edge

Every request passes through Cloudflare, which turns away attacks and floods of traffic before they reach the platform.

Watched around the clock

Outside checks call every service every minute, errors reach us as they happen, and the status page shows how each service is doing.

If this happens

What happens when something goes wrong

The things most likely to go wrong, and what you would notice.

The main server fails

Traffic moves to the standby, the database switches to the standby's copy in seconds, and scheduled work such as reminders and payment instalments carries on within a minute. When the main server is back, it catches up and takes over again.

CMS Orbit is down

Websites built with our SDK read the last published pages, menus and settings from a copy kept in Ireland, so visitors still see the site. Editors wait to publish until it's back.

One satellite is down

The others carry on. Each satellite has its own data and only talks to the rest through Orbit Exchange, so Booking Orbit going down doesn't stop CMS Orbit.

A message can't be delivered

When a satellite or your own system is briefly unreachable, the event waits and is sent again, with growing gaps, until it's confirmed. Nothing has to be sent by hand.

Someone deletes the wrong thing

Deleted pages, files and websites go to the bin and can be restored, and every page keeps its earlier versions. Only the Owner can erase something for good.

The service is attacked

Every request passes through Cloudflare first, which turns away attacks and floods of traffic before they reach the platform.

Commitments

What you can count on

What each case means for your data, and how we make sure it holds.

  • If a server fails

    What it means

    Seconds of data at most, because the standby has every change as it happens. Back within minutes, automatically.

    How we check

    We switch servers off in production on purpose, every quarter.

  • If a whole site were lost

    What it means

    Back from the most recent backup: at most an hour of changes.

    How we check

    Backups are restored on a separate machine and the apps started on them, every month.

  • Availability

    What it means

    We work to 99.9%, about 45 minutes a month. Enterprise customers can have it in their contract, with service credits.

    How we check

    Checked from outside every minute, and shown on the status page.

Drills

We break it on purpose

A plan that hasn't been tested is a hope. We switch things off in production on purpose, watch what happens and fix what we find.

  1. 01

    The main server switched off

    Under steady load, the standby answered every request without an error, and traffic was back on the main server within two minutes of switching it on again.

  2. 02

    The database moved to the standby

    The standby's copy took over in the same second. A publishing check made more than 2,000 changes meanwhile without an error, and found every page as it was last written.

  3. 03

    Backups restored from scratch

    Production backups were downloaded and restored on a separate machine in under a minute, with every record and index intact, and the apps started on them.

Data protection

Your data, written down

We use a few providers to run the service. What each one does, where and under which safeguards is on the sub-processors page.

Encrypted in transit

Every connection uses TLS, including the copy of the data between our servers.

Domain and email protected

DNSSEC signs our domain's answers so nobody can send you to a fake address, only certificate authorities we name may issue certificates for it, browsers are told to use HTTPS only, and DMARC sends email that pretends to be from us to spam.

Files, backups and secrets encrypted

Files and backups are encrypted in storage. Credentials for connected services, such as payment keys, are encrypted with AES-256-GCM under a key held in AWS Key Management Service, which only our servers can unlock and which logs every use. API keys and sign-in codes are stored only as hashes.

No tracking of your people

No analytics or advertising tracking of the people who use the service. Sign-in records are deleted after 12 months.

Sub-processors

Every provider, what they do for us and where they process data.

Privacy policy

What we collect, why, how long we keep it and your rights.

Data processing agreement

Our agreement under Art. 28 GDPR, with the measures we take.

Environment

The energy behind the platform

What our hosting provider, IONOS, reports for its own data centres, and how we keep our own share small.

Renewable electricity

IONOS reports that its own data centres have run on 100% renewable electricity for many years, and that 99.8% of the energy across its infrastructure was renewable in 2025.

Hardware reused

IONOS reports that all of its retired IT hardware was recycled or reused in 2025.

Little water for cooling

IONOS reports that its data centres use 99% less water than the industry average, because they avoid water-intensive cooling.

Waste recovered

IONOS reports that 98.9% of its data centres' waste was recycled, reused or recovered in 2025.

Shared servers

Every satellite runs on the same servers instead of its own, so adding a satellite adds little.

Served from cache

Much of the traffic is answered from cache, so it never reaches a server.

Accessibility

Made to work for everyone who uses it

We aim for WCAG 2.2 AA on the website, in the docs and in every app, in light and dark mode, and check every screen with automated tests in both.

CMS OrbitAccessibility

Continuous security testing

Bifröst Orbit takes security seriously, tests continuously and keeps improving. Security isn't a one-off exercise: we keep checking the platform for weaknesses and fix what we find.

How we test

  • Every week: an independent security service checks everything we expose to the internet, our websites, apps and APIs, for known vulnerabilities and weak settings.
  • Every change: each change to our code is checked for dependencies with known vulnerabilities and for secrets that don't belong in it, and the same checks run again every week.
  • Every finding: we review what the checks report, confirm whether it is real and fix what is, the most serious first, as part of our regular maintenance.

We believe security belongs in the way a platform is designed, built and run, not added afterwards. If you find something we missed, tell us below.

Report a security problem

If you find a vulnerability in Bifröst Orbit, email security@bifrostorbit.com. A person on our team reads every report. You can also use Report a vulnerability on any of our public repositories on GitHub, and our security.txt lists the same address.

What to send

  • What you found and where, with the steps to reproduce it.
  • What someone could do with it.
  • How to reach you, and whether you'd like to be credited.

While you test

  • Use your own account and organisation. Don't read, change or delete other organisations' data, and if you reach it by accident, stop and tell us.
  • Don't run denial-of-service or load tests, flood our forms or emails, or try to trick our people.
  • Give us a reasonable time to fix the problem before you tell anyone else.

What we do

  • Confirm that we have your report, and tell you what we decide.
  • Fix what we confirm, the most serious first, and tell you when it's fixed.
  • Credit you, if you'd like. We don't pay bounties.

We won't take legal action against research that follows these rules in good faith.

From the blog

More about security

All posts
CMS Orbit

4 min read

Reviews, approvals and an audit trail, built in

How CMS Orbit keeps content right: reviews on every plan, approval steps and review dates on Growth, approved claims, emergency publishing and a record of every change.

Questions about security?

Send us your security questionnaire, or tell us what your organisation needs to know, and we'll answer it.

Talk to us