Security you can check,
not just trust.
Every statement on this page matches how Valdra actually works. Where we have not done something yet, we say so.
Version 1.4 · 5 November 2026
The short version
The answers reviewers ask for first
Where is our data?
In Canada, in Beauharnois, Québec. Backups are in Canada too.
Is it encrypted?
TLS in transit, TLS 1.3 where the browser supports it. Uploaded files are encrypted with AES 256, with a separate key for each workspace. Backups are encrypted.
How do people sign in?
By emailed link, passkey or single sign on. There are no passwords. Your administrator can require a second factor for everyone.
Is access logged?
Yes. An append only audit log that you can export yourself.
Do you train AI on our data?
No.
Do you have a SOC 2 report?
Not yet. We say so plainly below.
Uptime commitment?
99.0% monthly, with service credits.
Breach notification?
Within 48 hours of us becoming aware.
Where your data lives
- Your records are stored in Canada, in Beauharnois, Québec, in our hosting provider's data centre. Encrypted backups are copied to a separate server in Canada.
- A few listed services process limited data outside Canada. Our Sub-processor List names each one and what it receives.
- We do not store primary customer data outside Canada, and will not without 30 days' notice and your right to terminate.
Personal information is removed before AI
- Valdra uses AI to draft recommendations, policies and reports. Before any text goes to an AI service, our own Shielk redaction engine removes the personal information in it.
- Valdra refuses to send text to an AI service unless it has been redacted or contains no customer data.
- Only the content needed for the request is sent, never your whole workspace.
- We do not train models on your data, and neither does our AI provider on commercial API traffic.
- If you use the Valdra AI gateway for your own tools, redaction is on by default and only your administrator can switch it off.
Encryption
- In transit: TLS on every connection, TLS 1.3 where the client supports it. Plain HTTP is redirected to HTTPS, and HSTS is enforced on every subdomain.
- Uploaded files: encrypted with AES 256 before they are stored, with a separate key for each workspace.
- Secrets: sign in secrets, signing keys and the access tokens of the services you connect are encrypted by the application. API keys we issue to you are stored only as a one way hash.
- Backups: encrypted before they leave the server. The key never travels with the offsite copy.
Sign in without passwords
- There is no password to steal and no password database to breach.
- People sign in with a single use link that expires after 15 minutes, with a passkey, or with your company's single sign on.
- Authenticator app codes are available to every account. Your administrator can require them for the whole organisation, and Valdra enforces that on our servers, not only in the browser.
- Repeated sign in attempts are rate limited.
Access controls
- Roles: admin, member, auditor and viewer, with least privilege by default.
- IP allowlists for sign in sessions and API keys, on the Professional plan and above.
- SAML or OIDC single sign on and SCIM user provisioning, on the Enterprise plan.
- On our side, production access is restricted to named individuals, key based only, with password sign in disabled.
- Any staff access to your workspace content, given under a support grant, is recorded in your audit log.
An audit trail you control
- The audit log records sign ins, administrative actions, changes to compliance records, and AI generated documents and drafts.
- Our database refuses to change its entries. Entries are removed only through a controlled erasure, when a workspace or account is erased.
- Each entry names the user and the time, and you can export the log yourself, without asking us. Regulators expect a custodian to produce its records independently.
Backups, monitoring and releases
- Encrypted backups every night, copied to a second server in Canada.
- A full restore of the offsite backup is tested every month.
- An automated check signs in and runs core workflows end to end every 30 minutes, and alerts us when it fails.
- Every release passes automated health checks and an end to end test. A release that fails them is rolled back automatically.
- Automated checks before each release block changes that break our privacy and security rules. Dependency vulnerability alerts are automated.
Error reports
- Errors in the application are reported to our error monitoring provider in the United States.
- Error reports never include request bodies or cookies. Before they leave our systems, a filter removes the personal details it recognizes, such as email addresses, phone numbers and postal codes.
Shielk, our redaction engine
- The engine learns from the shapes of what it detects, never from your text.
- How long Shielk keeps what you send is set out in its privacy policy.
If something goes wrong
- If your data is involved, we notify you within 48 hours of becoming aware, with what happened, what data was affected, the likely consequences and our response.
- That gives you what you need for your own report to the Privacy Commissioner of Canada or, where PHIPA applies, the Information and Privacy Commissioner of Ontario.
- Our uptime commitment is 99.0% a month, with service credits, set out in our Service Level Agreement.
What we do not claim yet
We do not yet have a SOC 2 report, and we are not certified to ISO/IEC 27001 or ISO/IEC 42001. A SOC 2 report is an attestation issued by a licensed CPA firm, not a certification.
We run our own information security and AI management programme inside Valdra, the same product you would be buying. We are a small company: [email protected] reaches the people who build the product, and we do not yet have 24/7 staffed operations or multi region redundancy. That is why our uptime commitment is 99.0% and not a number we could not keep.
If certification is a hard requirement for you today, tell us and we will be straight about the timeline.
Report a vulnerability
Found a security issue? Email [email protected] with the details and how to reproduce it. We read every report and reply.
Please give us a reasonable time to fix the issue before you share it, and do not access or change other customers' data while testing. Our contact is also published at /.well-known/security.txt.
Need a completed security questionnaire? We answer reasonable questionnaires at no charge.