CodeRunner

•
904
•

Executes AI-generated code in a secure local sandbox on your Mac using Apple containers.

CodeRunner acts as a Model Context Protocol (MCP) server, enabling remote Large Language Models (LLMs) like Claude or ChatGPT to run code directly on your local machine. This allows you to process sensitive local files (videos, images, documents, data) with AI without uploading them to the cloud, maintaining privacy and control. It provides a pre-configured, isolated environment with automatic dependency installation, a built-in skills system for common tasks, and seamless integration with various LLM platforms, transforming how you interact with AI-generated code.

Características Principales

01Secure sandboxed execution using Apple containers (VM-level isolation)
02Pre-configured environment with automatic tool and dependency installation
03641 GitHub stars
04Processes local files with remote LLMs without cloud upload
05Built-in and custom skills system for common tasks (e.g., PDF, image processing)
06Broad integration with LLM platforms (Claude Desktop, OpenAI Agents, Gemini-CLI, Kiro, Coderunner-UI)

Casos de Uso

01Execute AI-generated code from LLMs like Claude or ChatGPT locally to analyze, transform, or process files.
02Process sensitive local data (videos, images, documents) with remote LLMs without data ever leaving your machine.
03Integrate a secure code execution environment with various LLM frontends and tools for enhanced local AI capabilities.