Changelog

Structured root-cause analysis for Port State Control findings, built to support inspection follow-up and trend analysis across a fleet.

• Structured RCA workflow attached directly to PSC findings.
• Recurrence linking — connect related findings across separate inspections to surface repeat issues rather than treating each inspection in isolation.
• AI assistant scope extended to PSC queries, so questions about findings and root causes can be answered directly through chat.

A retrieval quality upgrade across the entire document corpus.

• Every document chunk now carries enriched context — vessel, document identity, and section topic — instead of being indexed as an isolated fragment of text.
• Search results are sharper because a match can be understood in context, not just matched on keywords in isolation.
• AI assistant answers that cite document content benefit the same way, since the underlying retrieval is the same context-enriched index.
• Applied across the full existing document corpus, not just newly uploaded documents.

Document ingestion gets faster and more resilient, especially for fleets onboarding a large existing archive.

• Bulk folder uploads — submit up to 50 files in a single batch instead of uploading one at a time.
• New cloud OCR lane for document processing, giving the pipeline more flexibility in how scanned documents are transcribed.
• Improved processing reliability, with better handling of edge cases in document ingestion.

A rebuilt permissions model that maps more accurately to how fleet organizations are actually structured.

• Three-tier role model: platform, company, and fleet level, replacing the previous flatter role set.
• Per-user vessel scoping — users can be limited to the specific vessels they're authorized to access, independent of their role tier.
• Route-level authorization hardening across the application, closing gaps between what the UI showed and what the API allowed.
• Admin panel improvements for managing roles and vessel assignments.

The foundational data model for fleet and crew management is now live.

• Vessel registry covering the core fleet data model.
• Rank structure configuration, aligned to standard maritime crewing hierarchies.
• Manning configuration for defining crew requirements per vessel.
• Notification platform to support alerts and updates as later features come online.
• Three-zone architecture: a shore-side office console, a seafarer-facing portal, and a public careers site, all built on the same underlying platform.

Vessels now generate structured periodic reports on a defined schedule, with a shore-side review workflow to match.

• Scheduled vessel reporting, configurable per report type and interval.
• Shore review workflow: submitted reports move through a clear review status before they're considered finalized.
• Report tagging, so reports can be organized and filtered by topic or category.
• Ask-AI on report content — query report data directly through the AI assistant instead of digging through individual submissions.