Cursor Chat History Vectorizer

Example Use Cases

• 1

  Search and retrieve relevant past chat prompts from Cursor IDE history.

• 2

  Perform vector similarity searches on local chat data for RAG or analysis.

• 3

  A simple API to access vectorized chat history stored in LanceDB.

Description

This project extracts user prompts from local Cursor IDE chat history, generates vector embeddings using a local Ollama instance, and stores them in a LanceDB vector database. It provides a Dockerized FastAPI server (MCP server) to perform vector similarity searches on the stored chat history, enabling easy retrieval and analysis of past conversations. This facilitates Retrieval Augmented Generation (RAG) and other LLM-based workflows by making chat history searchable and accessible through a simple API.