Thalassys

Open-source intelligence platform with hybrid FR/EN retrieval

Web platform In production

About the application

Thalassys continuously collects open sources - press articles, PDF reports, social posts - for each monitored domain, extracts the entities they mention and indexes everything for semantic retrieval. Analysts ask questions in plain French or English and get answers grounded in the collected corpus, every statement traced back to the document it came from.

The platform is multi-domain: one deployment monitors several watch areas at once, each with its own sources, topics and reporting cadence. It also produces its own analyst reports, from raw source summaries up to consolidated periodic assessments.

The first production deployment covers naval and maritime security. The same engine is domain-agnostic and can be pointed at any watch subject.

What it does

Automated multi-source collection (news, PDF, social)
Hybrid FR/EN semantic + keyword retrieval
Answers with per-document citations
Entity extraction and cross-domain graph
Automated periodic analyst reports
Multi-domain, multi-tenant deployment

Have an application to build?

Projects are quoted individually, after a short framing conversation about what the application has to do and who will use it.