Industry AI · AI assistant
VisionNova
Prototype
The assistant that sees what it is shown and answers what it is told.
Multimodal virtual assistant: it detects the objects present in a scene, interprets the inputs it receives and generates an answer in natural language.
How VisionNova works
VisionNova handles, in the same exchange, what it sees and what it is told: it detects the objects present in a scene, interprets the request received and formulates an answer in natural language.
The project is at prototype stage. It targets situations where the user shows an object rather than describing it: identifying a part, visual inspection, field support.
Where your data runs
- 0 data sent to a third-party service
- 0 token-based billing
- 100 % of the model executed on your hardware
The model is installed on your premises. An internet outage interrupts neither the assistant nor document retrieval. How we deploy AI on your own servers.
Technologies
Where it is deployed
The other AI in the range
-
DentalNova In production
The dental practice's administrative assistant, partitioned by role.
-
ProspectNova In production
Sales prospecting driven by your own prompts.
-
RecrutNova In production
Sourcing qualified profiles, described in natural language.
-
MedicalNova In production
The clinic's management assistant, partitioned department by department.
-
AssistNova In production
The website chatbot that answers from your real data, in real time.
-
StockNova In production
What is left, what has moved, what is about to run out — in one question.
Updated Aug. 11, 2026