Legal

Privacy Policy

What personal data NoFly processes, why, who else touches it, how long we keep it, and the rights you have over it.

Version 1.0 · Effective 9 August 2026 · Last Updated 9 August 2026

In plain termsSummary · the full text governs
  1. For account data we are the controller. For everything you scan, trace or route through NoFly we are a processor acting on your organisation's instructions - your admins decide, not us.
  2. Agent traces, prompts and scanned artifacts can contain personal data even though they are not designed to. We redact and minimise before storing, and the product will tell you where such data has reached.
  3. No advertising or third-party analytics trackers. We use browser storage for your session and preferences, not to follow you.
  4. Export and erasure are self-serve endpoints in the product. You do not have to file a ticket with us to exercise them.
Contents · 13 sections

1Who we are and what this covers

NoFly Security ("NoFly", "we") provides an AI-security platform. This policy explains how we handle personal data in the NoFly web application, API, command-line agent, SDK, browser extension and marketing site.

Registered address: Address to be confirmed - contact us for our registered details. Privacy contact: privacy@shomra.ai.

2What we collect

Account
Name, work email, hashed password, organisation and role, two-factor enrolment state, last-seen timestamp, and whether the account was provisioned by SSO or SCIM.
Organisation
Org name, membership and permissions, API-key metadata (a hash and a label - never the key itself), policies, integrations and audit records of governance actions.
Billing
Plan, entitlements and invoicing details for paid plans. Card details are handled by our payment provider; we do not store them.
Scanned content
Repository files, configuration, MCP server definitions, agent and skill specifications, model metadata, prompts, and anything else you submit for analysis.
Runtime telemetry
Agent traces and sessions, tool calls and their results, model requests and responses routed through our proxy or reported by the SDK, firewall decisions, and consumption figures.
Findings
The results of analysis: findings, evidence, verdicts, attack paths, red-team runs and evidence packs - with secrets redacted and raw bodies scrubbed before they are stored.
Product usage
Server logs, IP address, browser and device type, pages and actions, and error diagnostics, used to operate and secure the Service.
Support
Emails, contact-form submissions and anything you choose to send us when asking for help.

We do not knowingly seek special-category data (health, biometrics, beliefs, and similar). Where the product detects such identifiers it classifies them so you can act, and it deliberately excludes them from anything that becomes a reusable detection rule.

3Why we process it, and our legal basis

Providing the Service to your organisation and administering accounts
Performance of a contract
Processing the content you submit for scanning, testing and runtime control
Processing on the controller's instructions (Art. 28) - your organisation determines the basis
Securing the platform: authentication, abuse and fraud prevention, audit logging
Legitimate interests - running a secure service
Improving detection accuracy using de-identified, non-attributable signals
Legitimate interests - and we exclude personal data from this path
Billing, tax and corporate records
Legal obligation
Service notices, security advisories and product email you can turn off
Legitimate interests, or consent where the law requires it
Marketing email
Consent, withdrawable at any time

4AI processing

When AI augmentation is enabled, we send relevant content and metadata to our model provider for analysis. Three commitments apply:

  • Inference only. Our agreements with model providers prohibit using your content to train their models, and we do not train models on it either.
  • It is optional. AI augmentation can be switched off. With it off, the product runs its deterministic rule engine - a scan never fails because AI is unavailable.
  • Your content is treated as untrusted input. Anything we pass to a model is framed strictly as data; instructions embedded in it are handled as attempted injection, never as commands.

Automated analysis can produce a finding about a person's account or an agent identity, but it does not make decisions producing legal or similarly significant effects on an individual. Enforcement actions - blocking, revoking, suspending - follow policies your organisation configures, and a human can review and reverse them.

5Cookies and browser storage

We do not use advertising cookies, cross-site trackers or third-party analytics in the application. What we do use is strictly functional:

  • your session token, so you stay signed in;
  • your theme choice and the project or org you last had open, so the product opens where you left it;
  • security state such as two-factor verification and CSRF protection.

These are necessary for the Service to work, so they do not require consent. Clearing your browser storage signs you out and resets your preferences. If we later add analytics, we will update this section and ask for consent where the law requires it.

6Who else processes your data

We do not sell personal data, and we do not share it for advertising. We use a small set of subprocessors, listed with their purpose and region on our Trust & Security page:

  • a cloud provider for hosting and managed PostgreSQL, encrypted at rest;
  • a model provider for AI analysis, used only when AI augmentation is enabled;
  • a transactional email provider for account, invite and notification email;
  • a payment provider for paid plans.

We may also disclose data where the law compels us to, and will give you notice where we are lawfully able to. If we are acquired or merge, data may transfer to the successor under this policy, and we will tell you before it does. Each subprocessor is under a written contract limiting it to our instructions. We give advance notice of material subprocessor changes where your agreement requires it.

Anything you connect yourself - a code host, a ticketing tool, a chat workspace, a SIEM - receives what you configure it to receive, under your agreement with that provider.

7International transfers

We may process data in countries other than yours, including the United States. Where we transfer personal data out of the UK or EEA we rely on an adequacy decision where one exists, and otherwise on Standard Contractual Clauses with supplementary measures. Hosting region can be configured for some deployments - ask us before you sign if data residency is a requirement.

8How long we keep it

Account and organisation records
For the life of the account, then deleted or anonymised within 30 days of closure
Scans, findings and evidence packs
Until you delete them or close the account; deletions cascade to mirrored records
Agent traces and runtime telemetry
Per your plan's retention window, or until you delete them
Audit log
Retained for the account's life and after closure where we need it as a record of governance actions
Billing and tax records
As required by law, typically 6-7 years
Server and security logs
A rolling short window, typically no more than 90 days
Backups
Cycled out on the backup schedule after a deletion, typically within 35 days

Your org admins can delete most of this themselves at any time.

9How we protect it

Every query is scoped to a single organisation through a tenant guard that refuses to run unbound. Traffic is encrypted in transit and data is encrypted at rest. API keys are stored only as SHA-256 hashes. Untrusted code is analysed inside a disposable, network-isolated container - the application never executes what it scans. Access is role-based and enforced server-side, with two-factor authentication, SAML SSO and SCIM deprovisioning available. Governance actions are written to an append-only audit trail.

Our current compliance status - including which certifications we hold and which we do not yet hold - is stated plainly on the Trust & Security page. If a personal-data breach occurs, we will notify affected controllers without undue delay and support your own notification obligations.

10Your rights

Depending on where you live, you may have the right to access your personal data, correct it, delete it, restrict or object to processing, receive it in a portable format, and withdraw consent. Under US state privacy laws you may also have the right to know, to delete, to correct, and to opt out of sale or sharing - and we do not sell or share personal data as those laws define it.

For data we control - your account with us, our correspondence, marketing - contact privacy@shomra.ai. We respond within 30 days, and we will not charge you or make the Service worse for asking. If you are unhappy with our response you may complain to your local supervisory authority.

11Self-hosted deployments

If you run NoFly in your own environment, your data stays there. We do not receive your scans, findings, traces or user records, and this policy applies only to the limited data we hold about you as a customer - your licence, your contacts and your support correspondence. Similarly, the command-line agent performs local checks on your machine without sending source code to us.

12Children

The Service is for professional use and is not directed at anyone under 18. We do not knowingly collect data from children; if we learn we have, we delete it.

13Changes to this policy

We will update this policy as the product changes. The revision and date at the top of this page always reflect the current version, and we will notify you by email or in the product before a material change takes effect.

Questions, or a request for our Data Processing Agreement: privacy@shomra.ai. General enquiries: hello@shomra.ai or shomra.ai.

Reviewing NoFly for procurement? We can supply a Data Processing Agreement, our subprocessor list and current security documentation: email legal@shomra.ai.