Skip to content

Latest commit

 

History

History
97 lines (71 loc) · 2.56 KB

CORTEX.md

File metadata and controls

97 lines (71 loc) · 2.56 KB

Cortex CLI - Fork of Gemini CLI

This is a community fork of Google's Gemini CLI with enhanced features and additional AI provider support.

Key Changes from Original Gemini CLI

🔄 Rebranding

🆕 New Features

  • Ollama Support: Local AI inference with privacy-focused approach
  • Extended Provider Support: Framework ready for additional AI providers
  • Enhanced Authentication: CORTEX_AUTH_TYPE environment variable support

🔧 Technical Changes

  • Updated all package names and imports
  • Added OllamaContentGenerator class
  • Extended AuthType enum with USE_OLLAMA
  • Updated build system and bundle output

Installation & Usage

Install Cortex CLI

npm install -g @markcodering/cortex-cli

Basic Usage

# Interactive mode
cortex

# Direct prompt
cortex -p "Explain this code"

# With specific model
cortex -m llama2

Ollama Setup

# Install Ollama
curl https://ollama.ai/install.sh | sh

# Pull a model
ollama pull llama2

# Start Ollama server
ollama serve

# Use with Cortex CLI
export CORTEX_AUTH_TYPE=ollama
export OLLAMA_BASE_URL=http://localhost:11434  # optional
cortex -m llama2

Environment Variables

  • CORTEX_AUTH_TYPE: Set to 'ollama' for local AI
  • OLLAMA_BASE_URL: Ollama server URL (default: http://localhost:11434)
  • All original Gemini CLI environment variables are still supported

Migration from Gemini CLI

If you're migrating from the original Gemini CLI:

  1. Command Change: Replace gemini with cortex in your scripts
  2. Environment Variables: Optionally set CORTEX_AUTH_TYPE=ollama for local AI
  3. Configuration: Existing .gemini/ settings should work as-is

Original Features Preserved

All original Gemini CLI features are preserved:

  • OAuth authentication with Google
  • Gemini API key support
  • Vertex AI integration
  • MCP server support
  • File operations and tools
  • IDE integration
  • Checkpointing and memory

Contributing

We welcome contributions! This fork aims to:

  • Maintain compatibility with upstream Gemini CLI
  • Add support for additional AI providers
  • Enhance privacy and local AI capabilities
  • Provide community-driven improvements

Acknowledgments

This project is based on Google's excellent Gemini CLI. We're grateful to Google and the open-source community for creating such a powerful tool.

License

Apache 2.0 (same as original)