TrustSecurity Overview
Trust CenterQ1 2026

Security & Trust Center

PDFFillr handles KYC packets, fund subscription agreements, LP data, and regulatory filings. Here is exactly how we protect your documents β€” with full transparency on our controls, compliance posture, and incident response.

πŸ” AES-256 at Rest🌐 TLS 1.3 in TransitπŸ”‘ httpOnly Cookie AuthπŸ“‹ Immutable Audit Logs🚫 Zero AI TrainingπŸ‡ͺπŸ‡Ί GDPR CompliantπŸ‡ΊπŸ‡Έ CCPA CompliantπŸ“‹ NJ DLRA Compliant
Updated Q1 2026Jersey City, NJ, United StatesEngineersmind Corp

Our Commitment to Transparency

This page reflects what is actually implemented in our platform today, what is actively being improved, and what is on our roadmap. We will not claim controls we have not built.

01

Compliance & Certifications

EU

GDPR

Compliant

EU data protection β€” DPA, SCCs, and user rights framework in place

US

CCPA

Compliant

California Consumer Privacy Act β€” no data sale, rights requests supported per our Privacy Policy

NJ DLRA

Compliant

New Jersey Data breach notification obligations met

02

Our Security Commitment

PDFFillr is designed specifically for the security requirements of private equity and financial services firms. Our customers process KYC packets, fund subscription agreements, LP data, and regulatory filings through our platform. The controls described on this page reflect that responsibility.

Security is not an afterthought at PDFFillr β€” it is a foundational design principle that informs our infrastructure choices, authentication architecture, and data handling practices at every layer of the stack.

03

Authentication & Session Management

How users and systems authenticate to PDFFillr β€” verified from production code.

Platform users

Auth0 OIDC Β· JWT RS256 Β· verified via JWKS endpoint

Social login

Google OAuth via Auth0 β€” token never stored in browser

Token delivery

httpOnly cookies β€” inaccessible to JavaScript, not extractable by XSS

Access token expiry

15 minutes β€” short-lived by design

Refresh token expiry

30 days with automatic rotation on each use

Token replay detection

Family-based replay detection β€” stolen refresh tokens are revoked chain-wide

Machine-to-machine

Auth0 client-credentials flow for M2M integrations

API keys (developers)

Secret keys SHA-256 hashed before storage β€” never stored in plaintext

Cookie security flags

httpOnly: true Β· secure: true Β· sameSite: none Β· domain-scoped

httpOnly cookie delivery means your session cannot be stolen via JavaScript injection. Refresh token rotation with replay detection means a stolen token is detected and invalidated on its next use.

04

Encryption

How data is protected at rest and in transit.

Encryption in Transit

Live

All traffic between clients and our servers is protected with TLS 1.3. Auth0 communication enforces HTTPS at the SDK level. TLS 1.0 and 1.1 are not supported.

Encryption at Rest β€” S3

Live

All documents are stored in AWS S3 with server-side encryption (AES-256) enforced via bucket policy. No unencrypted writes are permitted at the bucket level.

Credential Hashing

Live

API secret keys and refresh tokens are hashed with SHA-256 before storage. Plaintext credentials are never written to the database.

05

Document Security

Controls specific to KYC, LP, and financial document processing.

Secure Document Storage

Live

Documents stored in AWS S3 with owner-scoped key structure: {orgId}/{uuid}/{filename}. No cross-tenant access β€” each organization's documents are logically isolated. Filenames sanitized β€” non-alphanumeric characters stripped before storage. No public S3 bucket access.

Presigned URL Access

Live

Documents are never exposed via permanent public URLs. All document access is via time-limited presigned S3 URLs generated per request. URLs can be obscured from browser history via backend proxy. Frontend does not upload directly to S3 β€” all uploads go through the authenticated backend.

No AI Training on Your Data

Guaranteed

PDFFillr does not use your documents, extracted data, or AI outputs to train, fine-tune, or improve any AI model. Documents are processed transiently within the AWS Bedrock boundary and are not retained for model development purposes.

Human Review Policy

Exception-only

Human access to your documents occurs only on an exceptional basis β€” for example, when you raise a support ticket or during a security investigation. All personnel with access are under NDA and confidentiality agreements. Human review is never routine or automated.

06

Audit Logging

Full traceability of document and API activity.

HTTP requests

All requests logged with method, URL, IP address, status code, and timestamp

Auth events

Token issuance, refresh, rotation, and replay attempts β€” all logged with user identity

Document events

Immutable DocumentAuditLog per document β€” upload, access, processing, deletion events

API key usage

Every API request updates last_used_at β€” usage is traceable per key

SDK requests

ApiRequestLog + UsageSummary models capture all programmatic API activity

Sensitive field redaction

authorization and cookie headers are automatically redacted from all logs

Log access

Log viewer protected by Basic Auth β€” not publicly accessible

Enterprise export

Audit logs exportable via API for SIEM integration (enterprise plans)

07

Infrastructure

Cloud provider

Amazon Web Services (AWS) β€” US regions

AI processing

AWS Bedrock β€” processed within AWS VPC boundary, no data leaves to external AI providers

Document storage

Private S3 buckets β€” no public ACLs, versioning enabled, AES-256 SSE enforced

Database

PostgreSQL β€” connection encrypted in transit via SSL

Rate limiting

Global: 15 requests / 60 seconds per IP β€” applied to all endpoints

Input validation

Global ValidationPipe with whitelist: true β€” unknown fields stripped and rejected on every request

File upload limit

10 MB per document β€” enforced at the server layer

Containerization

Docker β€” consistent, reproducible deployment environments

Secrets management

Environment-variable based β€” migration to AWS Secrets Manager in progress

08

Data Retention & Deletion

Data TypeRetentionDeletion
User documents (saved)Duration of active accountRemoved from active systems within 30 days of deletion request
Encrypted backupsUp to 90 daysOverwritten automatically in backup rotation β€” never actively accessed
Application logs90 daysRotated automatically β€” daily log files
API request logs90 daysPurged per retention schedule
Refresh tokens30 days from issuanceRevoked on logout, rotation, or replay detection
09

Incident Response & Breach Notification

Notification Commitments

  • Affected enterprise customers notified within 48 hours of confirmed breach
  • EU supervisory authority notified within 72 hours (GDPR requirement)
  • NJ DLRA notification obligations met
  • Notifications include: scope, data categories affected, timeline, and remediation steps

Response Process

0–1 hr

Automated alerting triggers incident response

1–4 hr

Affected systems isolated, access revoked, scope assessed

≀48 hr

Customer notification with initial impact assessment

Post-incident

Root cause analysis and remediation report issued

10

Vulnerability Disclosure

If you discover a security vulnerability in PDFFillr, please report it responsibly to Support@pdffillr.ai. We commit to acknowledging all reports within 48 hours and will work with you to resolve confirmed issues. We do not take legal action against good-faith security researchers.

All vulnerability reports are treated confidentially. We will notify you when a reported issue has been resolved and credit researchers who wish to be acknowledged.

Last UpdatedQ1 2026
Cloud ProviderAmazon Web Services
AI ProcessingAWS Bedrock (VPC boundary)
Security ContactSupport@pdffillr.ai