DrizzleTime/Foxel

What it solves

Foxel addresses the challenge of managing fragmented data across multiple cloud storage services and local drives. It provides a unified interface to organize, search, and automate interactions with files regardless of where they are physically stored.

How it works

Foxel acts as a central orchestration layer using an adapter-based architecture. It connects to various storage backends (like S3, WebDAV, Google Drive, and others) and exposes them through a single web interface. It integrates AI by using embedding providers and vector databases (like Milvus or Qdrant) to enable semantic search across unstructured data. Additionally, it features an AI agent that can perform file operations, web fetching, and processing tasks, with security enforced through role-based access control and an approval system for sensitive agent actions.

Who it’s for

This project is for individuals or teams who need a self-hosted, private cloud storage solution that integrates disparate storage providers and wants to leverage AI for intelligent file discovery and automated management.

Highlights

• Unified Storage: Centralizes local, cloud, and protocol-based storage (S3, WebDAV, SFTP) into one interface.
• Semantic Search: Uses AI embeddings and vector databases to find files based on natural language descriptions rather than just filenames.
• Integrated AI Agent: Includes an agent capable of web fetching and file system operations, with a secure approval workflow for write actions.
• Extensible Architecture: Supports a plugin system for custom routes and frontend components, plus a robust role-based access control (RBAC) system.
• Protocol Support: Exposes stored data via S3 and WebDAV protocols for compatibility with external applications.