Security & compliance

The same security on every plan

Every Norcube product is built with encryption and privacy as the default, not as a premium upgrade. The same protections apply to every user, from day one, at every scale.

Encrypted credentialsData stored in the EUGDPR compliantNo security tiers
Guarantees

What we will never do

These are engineering decisions built into the architecture, not policies that change with a new investor or quarterly target.

Never sell your data

Your data is yours. We don't monetize it, share it with anyone beyond the subprocessors that run the service, or use it for anything other than the service you signed up for.

Never profile you

No ad tracking, no profiling, no data brokers. In the dashboard we measure which features get used so we can improve the product. We never measure the content you store, and you can turn off the measurement too.

Never store credentials in plain text

Database passwords are encrypted with AES-256-GCM. API keys are hashed with HMAC-SHA256. This is the architecture, not a setting.

Security in practice

Credentials encrypted at rest

Sensitive credentials like database passwords are encrypted server-side using AES-256-GCM with AWS KMS for key management. They never exist in plain text in our storage.

API keys hashed, not encrypted

API keys are hashed with HMAC-SHA256 before storage, so we never hold the raw key. Each records its creation date and last use, and you can revoke it instantly.

Isolated backup execution

Every backup job runs on its own sandboxed server, created on demand. It decrypts credentials in memory, runs the backup and is destroyed. Jobs and customers share nothing.

Data stored in the EU

Your data is stored in AWS Frankfurt (eu-central-1). Status pages, CMS previews, SMS alerts and AI features use other providers, listed in the questions below.

Encrypted in transit

All API traffic goes through AWS Elastic Load Balancer with TLS encryption. Backup tasks connect to your database over TLS whenever your database offers it.

Encrypted at rest

Stored data is protected by AWS-managed server-side encryption. Client-side archive encryption for backups is on our roadmap as a high priority.

No request body logging

We don't log request bodies. Your credentials, translations, prompts and domain queries stay out of our service logs, which record routes, status codes and IDs.

CLI sessions you control

Every CLI sign-in is a session you can see and revoke under Account → CLI sessions. Refresh tokens rotate on every use, and a replayed token ends the whole session.

Least-privilege CI tokens

Automation uses organization CI tokens instead of personal sign-ins. Each carries only the scopes you grant, always expires and can be revoked at any time.

Monitor secrets encrypted per value

Monitor's secret headers, Slack and Discord URLs and webhook signing secrets are each encrypted with AES-256-GCM under their own data key from AWS KMS.

Access control

Members get per-product roles and resource-level policies, so you can give exactly what's needed. Two-factor authentication and SSO are on our roadmap, not built yet.

See the roadmap

Connections to addresses you choose

Monitor's checks and alerts, DomainRadar's lookups and page fetches and Backup's alerts all pass one guard, which refuses private addresses after DNS resolution.

Audit logs

Backup, Monitor, Norcube Auth and LangSync keep audit entries for a year, then delete them. Backup and LangSync show theirs in the dashboard.

Monitor data retention

Raw check results for at least 30 days, hourly rollups for 90 days, daily rollups for 2 years, resolved incidents for a year. The questions below have the full list.

Rate limiting

Sign-in and public lookup endpoints are rate-limited to stop brute-force attempts and abuse. Monitor also limits costly actions per user and per organization.

Compliance

Privacy & data protection

GDPR compliant

We process data lawfully, minimize what we collect, store it in the EU, and never sell it. Read the privacy policy.

Privacy policy

Customer DPA available

Our Data Processing Agreement is public and auto-applies to every customer. Signed copies available on request.

Read the DPA

Transparent subprocessors

Every third party that processes customer data is listed publicly, with what they see and where they are.

See the list

AI Act transparency

We use AI as a deployer, via OpenAI's API. LangSync labels AI translations, and your content never trains AI models.

On the certifications roadmap

We don't hold SOC 2 or ISO 27001 yet. We're building the controls first, and we'll pursue audits when customers ask for them.

Trust Center
Incident response

If something goes wrong

Detection and containment

We monitor infrastructure for anomalies. If a security incident is detected, affected systems are isolated immediately to prevent further exposure.

Customer notification within 48 hours

If a breach affects your data, we notify your organization's owners within 48 hours of confirming it, so you can meet your own 72-hour GDPR deadline. The notification says what happened, what data was affected and what we are doing about it.

Transparent post-mortem

After resolution, affected users receive a post-mortem explaining the root cause, timeline, impact, and what we changed to prevent recurrence.

If Monitor itself fails

Monitor's checks run from one place. If our checker is down, checks and alerts stop and you won't get alerted. Published status pages keep showing the last known state and say live updates are delayed. Our own health is at status.norcube.com.

Our status page

Report a concern

Report a vulnerability or security concern to [email protected] with the subject "Security". We reply within 5 business days, and sooner for anything urgent. Scanners can find this at /.well-known/security.txt.

The details

The longer answers behind the cards above.

Where exactly is my data stored?
In AWS Frankfurt (eu-central-1), with these exceptions: public status pages are stored and served by Cloudflare worldwide, CMS previews run on Fly.io in Frankfurt, SMS alerts go through BulkGate in the Czech Republic, and AI features use OpenAI in the US under Standard Contractual Clauses. Every provider is on our subprocessors page.
How do CI tokens limit what automation can do?
A token carries only the scopes you grant, never more than your own role. It always expires, can be revoked at any time and is stored only as a hash. Its prefix and checksum let secret scanners recognize a leaked token.
What happens if an encrypted Monitor secret is copied?
It won't open. Each value has a data key of its own, so a value copied to another record can't be decrypted there.
What does the connection guard stop?
It refuses private addresses after DNS resolution, so DNS rebinding and redirects into a private network don't get through. Monitor's checks also refuse sensitive ports such as SSH, SMTP and databases.
How long do you keep logs and audit entries?
Service logs are kept for 30 to 400 days depending on the service. Audit entries are kept for one year, then deleted automatically, unless needed for an ongoing investigation. When an Organization is closed, its audit entries are deleted with its other data. Our privacy policy has the details.
Can I read the audit logs?
Backup, Monitor, Norcube Auth and LangSync keep audit entries for a year, then delete them. Backup and LangSync show theirs in the dashboard. Monitor records every change made through its API, but you can't read those entries in the dashboard or the API yet.
What does Monitor keep, and for how long?
Raw check results for at least 30 days, removed a month at a time. Hourly rollups for 90 days and daily rollups for 2 years. Status changes and resolved incidents for a year. Alert deliveries for 90 days and usage counters for 400 days. Posts on a status page stay as long as the page. A subscriber's email address stays until they unsubscribe or the page or Organization is deleted, and an unconfirmed one goes after 7 days.
How do you use AI?
We use AI as a deployer, via OpenAI's API. LangSync labels AI-generated translations in the dashboard until a person edits them. Your content never trains AI models: we don't train models, and OpenAI doesn't use API data for training (it may keep requests up to 30 days to monitor abuse).

Questions about security?

We're happy to answer anything about how we protect your data. Reach out any time.

  • GDPR compliant
  • AES-256-GCM
  • HMAC-SHA256
  • AWS Frankfurt
// tick. tick. tick.
0 ticks since founding
Coming soon

Join the PromptHub waitlist

PromptHub is in active development. Leave your email and we'll invite you as soon as it opens, with a free tier and no credit card.

One email when access opens. No newsletter, no spam.