bionic-gpt/bionic-gpt

What it solves

Bionic provides a secure, self-hosted platform for organizations to build and deploy AI agents that interact with internal data and systems. It addresses the need for a controlled, private environment where AI models can execute code, access files, and perform business-specific tasks without relying on external, public-facing AI services.

How it works

Bionic uses a Unix-inspired virtual filesystem to manage agent interactions. Instead of loading all data into a model's context, the system exposes memory, datasets, skills, and tools as files and directories. This allows the AI to navigate, read, and manipulate resources using standard operations. The platform is built in Rust and designed for Kubernetes, supporting local, private, or hosted model backends while maintaining strict governance, audit logs, and team-based permissions.

Who it’s for

This project is designed for internal AI engineering and platform teams that need to integrate AI workflows into their existing enterprise infrastructure while maintaining data sovereignty and security.

Highlights

• Agentic Filesystem: Uses a directory-tree abstraction to organize memory, tools, and knowledge for agents.
• Model Agnostic: Supports a wide range of model backends, including local, private, and hosted options.
• Enterprise-Ready: Includes built-in support for SSO/OIDC, audit trails, role-based access control, and Kubernetes-native deployment.
• Extensible Runtime: Provides a sandboxed environment for code execution, tool usage, and reusable domain-specific "skills."