Built for the security review.

Encryption

Data is encrypted everywhere it lives and everywhere it moves.

At rest

AES-256 across the primary database, file storage, and backups.

In transit

TLS 1.3 for all client connections. Older protocols disabled.

Application secrets

Stored in a managed secrets store; never committed to source control.

Access control

Least privilege by default.

MFA

TOTP-based 2FA mandatory for all users. Backup codes issued at enrolment.

RBAC

Predefined Admin / Editor / Viewer roles plus custom roles per organization.

Audit log

Guide creation, edits, shares, deletes, and access changes are logged with actor and timestamp.

Your data, your AI, your bucket

Enterprise customers can keep the AI and storage entirely on their own accounts — we never see the bytes.

Bring your own AI provider

Plug your own Anthropic, Azure OpenAI, Google Gemini, or Snowflake Cortex credentials into Organization Settings. All AI calls (step descriptions, AI Enhance, workflow summaries, Ask Insta) route through your account, on your billing, under your data-processing agreement.

Bring your own bucket

Point new screenshots and audio at your own AWS S3, Azure Blob, or GCS bucket. We store the URL; the bytes never sit on our infrastructure.

Sensitive data redaction

Sensitive values get blurred from screenshots automatically — before any step is saved.

Automatic detection

Every captured screenshot is scanned for emails, phone numbers, US SSNs, driver's licenses, credit cards, IBAN / bank accounts, API keys (Stripe, AWS, GitHub, Google OAuth, OpenAI, Anthropic), JWTs, private keys, and secrets in URLs.

Granular control

Org admins pick which categories to scan from a single panel — by group (Contact info, Government IDs, Financial, Credentials) or by individual detector. Off by default; opt-in.

Two modes

Permanent (the original is replaced — everyone sees the same redacted image, safest for public guides) or Permission-based (the original is kept and revealed only to viewers with the View original role permission).

Audit-safe by design

Each step records what was redacted as a category list with masked previews. Raw values never leave the detection pipeline; audit logs and exports both show categories only.

Manual blur still available

Authors can still add manual blur regions via the annotation tool to redact anything the auto-detector missed.

Application security

The boring fundamentals, taken care of.

Dependency scanning

Automated scans flag known vulnerabilities in third-party packages; criticals patched on a fast track.

Code review

Every production change goes through code review before deploy.