Admin — Workspace

Deletes all vectors, conversation summaries, and cached metadata for the workspace. This is not model fine-tuning — it is RAG memory that can be safely wiped and rebuilt by re-uploading datasets.

About Memory & RAG

Easy Reports uses Retrieval-Augmented Generation (RAG) to “learn” from your data. When you upload a dataset, its profile and metadata are embedded into a vector store. When you ask questions, relevant context is retrieved to ground the AI's answers.

This is not model fine-tuning. The LLM itself is not modified. Memory is instant to build, cheap to maintain, and can be wiped at any time.