Safe Vibes has been introduced—an open-source Django-based framework designed for the secure deployment of AI-generated tools. The platform allows users to generate HTML and SQL reports via an AI interface, while ensuring strict access control and security policies managed by the IT department.
What Happened
Safe Vibes has been developed, an open-source Django-based solution for creating and deploying AI-generated reports. Users can interact with an AI builder to create HTML + SQL reports by connecting approved databases such as Postgres, BigQuery, and Snowflake. The system includes governance mechanisms, such as read-only SQL policy verification, limits on the volume of rows and bytes transferred, and access control via IP whitelists and SSO.
Context
The project focuses on creating the infrastructural wrapper for the secure execution of AI-generated code. It offers a ready-made template for creating isolated sandboxes for AI agents working with corporate databases, acting as a protective layer between LLMs and critical data.
Why It Matters for the Industry
The tool helps address the problem of "shadow IT," where employees use AI to create fragmented and insecure analytical tools that bypass corporate standards. Safe Vibes provides the necessary governance layer to legitimize the "vibe-coding" process at scale within companies, creating a controlled environment for AI agents.
Why It Matters for Users
Business teams gain the ability to quickly create tools through AI chat (vibe-coding), while IT departments maintain full control over data security, API costs, and compliance with internal access policies.
What Is Not Yet Known / Limitations
Currently, there is a divergence in positioning approaches: ranging from a purely infrastructural solution (Django/security) to a product aimed at democratizing "vibe-coding."
Sources
Author
Look at AI, Editorial Team