Claude Code & Codex Provider Switcher addresses the need for managing and switching between different API configurations for Claude Code and Codex. It targets developers who frequently switch between various AI models or providers, offering a streamlined desktop application for this purpose. The primary benefit is simplified configuration management, saving time and reducing errors associated with manual switching.

How It Works

The application is built using Tauri 2.0, leveraging Rust for the backend and React/TypeScript with Vite for the frontend. This stack provides native performance and a small application footprint. The core functionality involves managing provider configurations stored locally in ~/.cc-switch/config.json. When a user selects a provider, the application copies the corresponding configuration files (e.g., settings-<name>.json for Claude, auth-<name>.json and config-<name>.toml for Codex) to the main application configuration directories (~/.claude/ and ~/.codex/). This strategy ensures that only the active configuration is present in the main directories, minimizing potential conflicts and simplifying the switching process.

Quick Start & Requirements

• Installation: Download the appropriate installer (.msi or .zip for Windows, .zip for macOS, .deb for Linux) from the Releases page.
• System Requirements: Windows 10+, macOS 10.15+, Ubuntu 20.04+/Debian 11+/Fedora 34+.
• Setup: Minimal setup time, as it's a desktop application. No specific hardware or API keys are required for installation, only for using the configured providers.

Highlighted Details

• Migrated from Electron to Tauri 2.0, resulting in an 85% reduction in application size (~12MB) and a 10x increase in startup speed.
• Supports both Claude Code and Codex, with presets for Qwen coder, Kimi k2, Zhipu GLM, DeepSeek v3.1, and packycode, as well as custom provider support.
• Information is stored locally in ~/.cc-switch/config.json, ensuring no privacy risks.
• Includes a "Official Login" option for both Claude and Codex to revert to default authentication methods.

Maintenance & Community

• The project is licensed under the MIT license.
• Contributions via Issues and Pull Requests are welcomed.

Licensing & Compatibility

• MIT License, which is permissive and generally compatible with commercial use and closed-source projects.

Limitations & Caveats

• macOS users may encounter a "unknown developer" warning on first launch due to the author not having an Apple Developer ID, requiring manual override in system security settings.
• The application manages API key configurations; it does not provide the AI models themselves. Users must have their own API keys for the respective services.