Your sensitive data is scattered, and you can't see it
PII lives in databases, S3 buckets, file shares, and CSVs nobody remembers. Pelestra scans every source in-place with context-aware detection to surface real findings, not noise.
Applied AI Workflow Layer - Data Readiness
Pelestra is Dotnitron's data readiness workflow layer. We use it to scan, classify, and protect sensitive data across your infrastructure before deploying any AI model.
Repository Review
Context-Aware Detection
Readiness Assessment
Read-Only Scanning
The Problem
Sensitive data scattered across databases, cloud storage, and file systems with no centralized visibility
Regex-only detection flooding teams with false positives while missing context-dependent PII
Manual audits producing stale snapshots that are outdated before they are finished
Cross-border data transfers happening without detection or lineage tracking
Platform Capabilities
PII lives in databases, S3 buckets, file shares, and CSVs nobody remembers. Pelestra scans every source in-place with context-aware detection to surface real findings, not noise.
Multi-tier scanning: pattern matching, NLP-based named entity recognition, and exact data match. Reads column headers and surrounding context to reduce noisy findings.
PostgreSQL, S3, Azure Blob, local filesystems, and more. Read-only, agentless connections keep source data inside the approved environment.
Private Docker delivery with licensing controls. SSO, RBAC, and audit trails are designed around your perimeter and operating rules.
Track how sensitive data flows between systems. Detect cross-border transfer risks and visualize lineage patterns, turning discovery into actionable risk mitigation.
How It Works
Agentless connectors for PostgreSQL, S3, Azure Blob, local file systems, and more. Read-only access keeps source data in place.
Define scan policies with glob patterns and get PII findings with confidence scores and surrounding context.
Monitor discovery volume and PII density across sources in real time. Track lineage before sensitive data spreads.
Detection Engine
Regex patterns detect structured PII such as credit cards, phone numbers, and national IDs. Fast, broad, and configurable per jurisdiction.
Named entity recognition identifies context-dependent PII such as names, addresses, and organizations that regex alone can miss.
Compare against known datasets for stronger confirmation. Column headers and surrounding context help validate findings.
Enterprise Security
PostgreSQL, S3, Azure Blob, and file systems with read-only access
Regex, NLP, and exact data match with column-header context
Data is streamed and chunked in memory so raw PII does not need to be stored
Track sensitive data flows and detect cross-border transfers
Board-ready PDF and CSV exports of findings and asset matrices
Scans, findings, and actions can be logged with provenance for review
Per-engagement enforcement via secure licensing bricks
SAML, OIDC, and role-appropriate access control
Deployment
Your data center, your rules
Secure delivery for advisory and audit firms
OpenAI partner network
Ready for a production decision