Chat With Your Files And Database Locally, Open Source

Assistant Engine is a C#/.NET 9 AI interface that runs locally, ensuring data privacy. It allows developers to orchestrate local or remote models through Ollama. The engine provides features like C#-aware reasoning, Retrieval-Augmented Generation (RAG), PowerShell execution, and natural language to SQL translation. It employs a multi-model approach with specialized roles such as Assistant, Embedding, Descriptor, and Text-to-SQL. The RAG pipeline transforms data into a searchable format through description, chunking, vectorization, and storage in local SQLite vector stores. Users can point the UI to files or databases for ingestion and can query them using natural language. The engine supports SQL Server, Postgres, and MySQL databases and can generate schema summaries. Built entirely in .NET 9/C#, it operates without cloud or Python dependencies, offering complete control over speed, cost, and accuracy. Documentation is available for RAG Ingestion, Database Access, and Tools.