BoCloud/folib

What it solves

FOLib is a comprehensive software supply chain service platform designed specifically for AI research and development. It addresses the need for a centralized, multi-language artifact repository that can handle both traditional software packages and modern AI model weights, while integrating security and intelligence into the supply chain process.

How it works

FOLib acts as a unified repository manager that supports over 23 different language ecosystems (such as PyPi, Maven, npm, and Docker). For AI specifically, it provides proxy and synchronization capabilities for major model hubs like Hugging Face, Ollama, and ModelScope, allowing teams to host private versions of these models. It utilizes a graph-based data model to track relationships between requirements, services, artifacts, security vulnerabilities, and dependency certificates. Additionally, it implements the Model Context Protocol (MCP) to enable AI agents to perform intelligent queries, recommend artifacts, and automate vulnerability patching.

Who it’s for

It is designed for AI researchers and software engineers who need to manage a complex supply chain of dependencies, container images, and AI models across multiple programming languages in a secure, private, or cloud-native environment.

Highlights

• Extensive Language Support: Compatible with 23+ package managers including PyPi, Maven, Docker, Cargo, and Conda.
• AI Model Integration: Native proxying and synchronization for Hugging Face, Ollama, and ModelScope.
• Agentic Capabilities: Support for MCP and AI agents to automate artifact discovery and security remediation.
• Cloud-Native Ready: Supports OCI image formats, multiple container clients (podman, nerdctl, etc.), and WebDAV for large file mounting.