We’re open to suggestions on the specific stack — but whatever we build must be production-ready from day one.
The priority is a robust, reliable backbone for data ingestion, tagging, and retrieval — built with stability, traceability, and maintainability in mind. The architecture will be Docker-based, running a lightweight instance on the client’s internal network, where it securely connects to local file shares, SQL databases, and other sources. This backend processes and indexes the data locally, with a secure connection to our multi-tenant cloud front end, where users interact with the system via the ViewOps interface. No raw data leaves the client environment — only safe, structured outputs used to power search, tagging, and AI-driven queries.