Privacy Policy

Shieldly is built on a simple principle: your AWS policy content is never stored. We analyze it, return the results, and discard it. The only data we retain is anonymized service-usage metadata required to operate your subscription and usage dashboard.

1. You paste or upload a policy in your browser.
2. We compute a SHA-256 hash of the input for cache deduplication. The hash is one-way and mathematically irreversible.
3. If no cached result exists, the policy text is sent over TLS to an AI model API for analysis.
4. The AI response (findings, recommendations) is returned to you.
5. The policy text is discarded. It is never written to any database.
6. The AI findings are also discarded — unless you explicitly enable History or use the Share feature.

We store SHA-256 hashes to identify duplicate analysis requests so we can serve cached results without re-charging you. A SHA-256 hash cannot be reversed to reconstruct your input. The hash reveals nothing about the content of your policy. Cache keys expire after 7 days.

If you enable History (the default for signed-in users), we store per-analysis metadata: timestamp, character count, units consumed, plan tier. We do not store the policy text or the full AI findings in the history record — only the score and severity summary. You can disable history at any time in Settings → Privacy.

When you click "Share Analysis", we store the AI-generated findings (risk summary, findings list, remediation suggestions) — never your original policy text. Shared links are public to anyone with the URL, expire after 30 days, and can be deleted immediately from your History panel. Link sharing can be disabled in Settings → Privacy.

Your policy text is transmitted to one of our AI providers (under enterprise API agreements) for inference. These providers are contractually prohibited from using API data for model training. The data is processed transiently and not retained after the response is delivered. We rotate AI providers based on your plan tier.

Sign-in and account management are handled by Clerk. Clerk stores your email address, name, and OAuth tokens. We receive only an opaque user ID from Clerk — we do not store your email or name in our own databases unless you explicitly set them.

Billing is handled entirely by Lemon Squeezy, which is PCI DSS Level 1 certified. We never receive, process, or store credit card numbers or billing addresses. We receive only a webhook notification confirming your subscription status.

If you use the AWS Connect feature, you create a read-only IAM role in your account and provide us the role ARN and an ExternalId. We store only the role ARN and ExternalId (encrypted at rest) to perform on-demand policy scans. We never store AWS credentials. Connections can be removed at any time from the AWS Account tab.

We use PostHog and Google Analytics to understand feature usage. Events are anonymized — they include action names ("analysis_run", "plan_upgrade_clicked") and plan tier, but no personally identifiable information. You can opt out of analytics by enabling a browser-level do-not-track signal.

• Settings → Privacy → Save analysis history: Toggle off to stop storing usage metadata going forward.
• Settings → Privacy → Allow link sharing: Toggle off to hide the Share button and block share creation.
• Profile → Delete Account: Removes your user record and all associated metadata within 30 days.
• Email request: privacy@shieldly.io for data access, correction, or deletion requests.

All data is encrypted in transit (TLS 1.3) and at rest (AES-256 via AWS DynamoDB encryption). API keys are stored as one-way SHA-256 hashes — the plaintext is shown once at creation and never stored. AWS infrastructure is hosted in us-east-1 with SOC 2 and ISO 27001 compliance.

We will notify signed-in users by email at least 14 days before any material change to this policy. The current version is always available at shieldly.io/privacy.

Privacy questions: privacy@shieldly.io
General support: support@shieldly.io