Conceptual design for NoSQL databases, with AI-assisted translation
NoER (Not Only ER) starts from a simple observation: document databases like MongoDB have too often turned data modeling into a purely application-driven exercise, skipping the conceptual phase that’s standard practice in relational databases. NoER brings conceptual design back into the NoSQL world, using an extended version of the Entity-Relationship model (called ERd) built specifically for document-native structures.
The tool guides users from a conceptual schema all the way to a JSON schema ready for MongoDB Atlas, through a translation assistant that applies document design patterns (embedding, referencing, extended referencing, subset) based on cardinalities and participation constraints.
Three translation modes, from manual to fully AI-guided
- Manual pattern selection — for each relationship, choose among the MongoDB patterns the system offers.
- Guided selection — NoER suggests a pattern based on data access and update frequency, along with each option’s pros and cons.
- AI-driven translation — NoER generates a complete translation plan for the whole schema, with rationale and confidence level for every relationship. Integrated models include Claude Haiku 4.5.
Where it comes from
NoER is the natural evolution of designER into the NoSQL world, built with the same research group at Politecnico di Torino (Andrea Avignone, Silvia Chiusano, Sara Giorza, Santa Panduri) together with Riccardo Torlone (Università Roma Tre). It’s already used in a Master’s course in Computer Engineering at Politecnico to introduce students with a relational background to NoSQL modeling.