Prospero (local LLM infrastructure)

TITLE: Prospero (local LLM infrastructure)

ABSTRACT: Prospero is a private, locally hosted AI research assistant built to make personal knowledge bases genuinely useful. The system combines local model routing, document-grounded retrieval across private collections, persistent memory, collection management, diagnostics, and backup/export tooling into a single browser-based workbench. Built largely through iterative “vibe-coding,” Prospero is nonetheless a practical and increasingly capable platform: it can ingest local folders of documents, search across private RAG collections with metadata filters, route general and coding tasks to different local models, remember explicitly saved user/project context, and preserve user control over what is stored, searched, and exposed. Its value lies in demonstrating how a local-first AI assistant can become a real research companion: private, extensible, inspectable, and tailored to the way its user actually works.