Nexus Local
Local-first desktop AI assistant for private knowledge retrieval and workflow support.
Nexus Local is a privacy-focused desktop assistant built to run on local infrastructure. It combines document ingestion, retrieval-augmented generation, workspace-aware knowledge access, and agent-style task routing so users can query private files without sending data to external AI services by default. The project is designed around a native desktop workflow rather than a browser-first chat interface.
Context
Problem Solved
Many AI tools are useful but unsuitable for sensitive notes, internal documents, research material, or codebases because they rely on external APIs and cloud storage. Nexus Local explores a more private alternative by keeping retrieval, storage, and orchestration on the user's machine while still providing a practical assistant experience.
Execution
Key Features
Native desktop chat interface built with Flet
Local document ingestion and retrieval using LanceDB
Workspace-scoped knowledge access for private files
LangGraph-based routing across local retrieval, web lookup, identity, and tabular analysis tasks
SQLite-backed chat history and workspace state
Optional web search when a query needs external context
Sandboxed tabular analysis for CSV and spreadsheet-style questions
Background file watching for workspace ingestion and local organization flows
Build
Tech Stack
Links
Explore the Project
Gallery
Product Gallery
A visual walkthrough of the interface, workflows, and key states captured from the project.
