How ScreenJournal protects your data

Updated on 19 July 2026

ScreenJournal is built so that the most sensitive thing it touches — a live view of your screen — is the thing it keeps for the shortest time. The work is read as it happens, turned into a timeline, and the raw screen data is deleted during processing. On top of that privacy-first foundation, everything ScreenJournal does store is encrypted, access is held to least privilege, and the whole platform is monitored continuously. This page explains how.

The most private data is the one we never keep

ScreenJournal's default is derive-and-discard: the screen is captured as short-lived video, a frontier AI model reads the work from it, and the video is deleted immediately during processing. What survives is understanding — plain-English entries of what was done, in which app, for how long — not footage. There is no screenshot gallery to scroll, no video to replay, and nothing for an attacker or an insider to exfiltrate.

The one deliberate exception is Evidence Mode, an optional add-on a company can switch on for the specific roles it flags as a data-leak risk. It is off by default and disclosed to the affected employee in the product. Everyone else stays on the derive-and-discard default.

How your data is encrypted

In transit. Every connection to ScreenJournal is encrypted with TLS (HTTPS), reaching us through Cloudflare's global edge, which also provides DNS and DDoS mitigation.

At rest. Everything we store is encrypted at rest using AWS Key Management Service (KMS):

• The primary Amazon RDS for PostgreSQL database is encrypted with a KMS key.
• Every Amazon S3 bucket uses server-side encryption and has S3 Block Public Access fully enabled, so no object is ever reachable anonymously.
• Backups and their cross-region replicas are encrypted with the same discipline.

Where ScreenJournal runs

ScreenJournal is hosted on Amazon Web Services in the ap-south-1 (Mumbai) region, on managed services rather than hand-run servers:

• The application runs on an auto-scaling group of EC2 instances behind an Application Load Balancer, released through AWS CodeDeploy for controlled, reversible deployments.
• The database is Amazon RDS for PostgreSQL, deployed Multi-AZ for resilience, encrypted, not publicly accessible, with 14-day automated encrypted backups and deletion protection enabled.
• Network access to the database is restricted to the application tier — it does not accept connections from the public internet.
• Interactive SSH to production servers is disabled; there is no open port 22.

Who can access what

Access to ScreenJournal — both your account and our systems — follows least privilege.

Your account. ScreenJournal accounts are passwordless: you sign in with a magic link or a connected account rather than a reusable password. You can add two-factor authentication with any authenticator app, backed by downloadable one-time recovery codes. From Settings → Sessions you can see every device signed into your account, with its IP address and last-active time, and revoke any of them or sign out everywhere else. If an administrator ever impersonates your account for support, that session is flagged in your list with who did it — there is no hidden access.

Our systems. AWS access is scoped to the specific resources each role needs. The AWS root account is held by a single custodian, is not used for day-to-day work, and any attempt to use it is alerted on through AWS CloudTrail and Vanta.

How we build and ship securely

• Changes are peer-reviewed and must pass automated tests before they can be promoted to production; production branches are protected against unreviewed direct pushes.
• Every repository runs a recurring, multi-stage security review — an automated AI security pass, an OWASP Top 10 scan, and engineer peer review — with findings and their remediation tracked in each repository.
• All repositories have GitHub Dependabot and GitHub-native security scanning enabled for continuous alerts on vulnerable dependencies.
• Company laptops are continuously checked by Vanta Device Monitor for disk encryption, screen lock, OS patch level, and active anti-malware.

Data retention and subprocessors

Derived data is kept only as long as it is useful and then expires automatically:

To turn screen and audio into a timeline, ScreenJournal sends captured content over HTTPS to two third-party subprocessors: Google Gemini (activity-timeline generation) and DeepInfra (speech-to-text transcription). Transcription is not self-hosted, and neither subprocessor is used to train models on your data.

Monitoring and incident response

The platform is monitored continuously. Operational and security signals flow through Amazon CloudWatch and AWS CloudTrail (multi-region, tamper-evident audit logging of every account action), plus edge and endpoint health checks and Vanta's continuous compliance monitoring. Critical alerts — a failed health check or a security event — page a 24/7 on-call engineer through PagerDuty, so there is a defined, accountable path from detection to response.

Compliance

ScreenJournal is actively pursuing SOC 2 Type II and ISO/IEC 27001, with its controls managed and continuously monitored through Vanta. We are happy to walk prospective customers through our current status and share evidence under NDA.

For how we handle personal data — PII removal during processing, employee redaction, and what a manager can and cannot see — read we record work, not people and our privacy policy.

Reporting a security issue

If you believe you have found a security vulnerability in ScreenJournal, email security@screenjournal.ai. We welcome responsible disclosure and will work with you to confirm and resolve the issue.

Frequently asked questions

Is my screen recorded and stored?

No. By default ScreenJournal reads your on-screen work to derive a timeline of what was done, then deletes the raw screen data during processing. There is no screenshot or video archive to store, watch back, or leak. The one exception is Evidence Mode, an optional add-on a company can enable for specific flagged roles, which is disclosed to the affected employee in the product.

How is my data encrypted?

Every connection to ScreenJournal is encrypted in transit with TLS (HTTPS). Data stored on our side is encrypted at rest with AWS-managed KMS keys: the PostgreSQL database is encrypted, and every S3 storage bucket uses server-side encryption with public access fully blocked.

Where is my data hosted?

ScreenJournal runs on Amazon Web Services in the ap-south-1 (Mumbai) region. The database is Amazon RDS for PostgreSQL, deployed Multi-AZ for resilience, not publicly accessible, and backed up automatically with 14-day encrypted retention.

Is ScreenJournal SOC 2 or ISO 27001 certified?

ScreenJournal is actively pursuing SOC 2 Type II and ISO/IEC 27001, with its controls managed and continuously monitored through Vanta. We are happy to share our current status with prospective customers under NDA — contact security@screenjournal.ai.