Devana is an AI operating layer from Scriptor-Artis for organisations in regulated sectors. Through a visual no-code environment, teams design AI workflows, connect data sources and deploy agents with their own permissions. A proprietary engine, ODIN, processes documents in numerous formats, and every AI call is recorded in an audit trail with source citations. According to the vendor, Devana runs on premise, in a data centre, at the edge or as a sealed appliance, including without an internet connection.
Scriptor-Artis SAS is a French company with its registered office in Mazères; no parent company is named. Devana can run on premise, so data can stay in your own environment; the cloud region is not stated. The privacy page says Devana may use external model providers as sub-processors for inference, including Scaleway, OpenAI, Anthropic, Google, Mistral AI and Meta, depending on what is enabled; some may be outside the EU, with standard contractual clauses where required. It does not say which are active by default: define which models and locations are permitted.