Vannus / Catalog / LightOn

LightOn C+ Moderate

Graded C+ on resilience. In Vannus's assessment it has not disclosed which model it runs — acceptable, with limits worth knowing.

Production multimodal RAG API from LightOn SA, 2 rue de la Bourse, 75002 Paris, France (RCS Paris 821 100 690). Three endpoints: /parse for document parsing and OCR via LightOnOCR-2 across 20+ languages, /extract for schema-driven structured field extraction, and /search for hybrid retrieval with citations. Answers are grounded to source passages by default and the platform is agent-native via MCP

AI, research, enterprise
Model provenance

Not disclosed. LightOn does not publish which model it runs, so this criterion is excluded and the grade rests on what we could verify. We say so rather than guess.

Who controls it
Allied / Safe Harbor Origin
OriginFRA
Data jurisdictionUS
Within U.S. CLOUD Act reachNo
Trains on your dataYes, unless you opt out

This is a separate question from the grade above. The grade measures resilience — whether the tool endures and whether you could leave it. This measures who can compel your data. A tool can score modestly on one and strongly on the other, and many do.

Signals on file
Origin
France · data in United States
US legal control
No
Training on your data
Trains on your data unless you opt out

The vendor's published or catalog-recorded posture — the concrete facts this grade is built from. A full audit verifies each against the vendor's current documentation.

How this grade is set

Vannus grades LightOn against nine dimensions of trust — data sovereignty, training privacy, compliance posture, operational resilience, exit portability, and more. The heaviest criterion asks whether the tool builds its own AI or resells someone else's model; where the vendor discloses it, the grade cites the vendor's own documentation. No paid placements — scoring is walled off from affiliate revenue. See the methodology →

Related tools we grade
Visit LightOn ↗ Grade your whole stack →