opencursor 0.0.5

An AI-powered code agent for workspace operations

OpenCursor

An AI-powered code agent for workspace operations.

Features

• Chat with an AI coding agent in autonomous or interactive mode
• Direct LLM chat without tools
• File context management (add, drop, clear)
• Repository mapping
• Focus on specific files
• Workspace directory selection

Installation

Using Poetry (recommended)

# Clone the repository
git clone https://github.com/yourusername/opencursor.git
cd opencursor

# Install with Poetry
poetry install

Using pip

pip install git+https://github.com/yourusername/opencursor.git

Usage

Once installed, you can use OpenCursor from the command line:

# Basic usage
opencursor -q "Create a simple Flask app"

# Specify a workspace directory
opencursor -w /path/to/workspace -q "Fix the bug in app.py"

# Use a different model
opencursor -m "gpt-4" -q "Refactor the authentication module"

# Run in interactive mode
opencursor -i -q "Create a React component"

Command-line Options

• -w, --workspace: Path to the workspace directory (default: current directory)
• -q, --query: Query to process (required)
• -m, --model: LLM model to use (default: qwen3_14b_q6k:latest)
• -H, --host: Ollama API host URL (default: http://192.168.170.76:11434)
• -i, --interactive: Run in interactive mode (one tool call at a time)

Development

Setup

# Clone the repository
git clone https://github.com/yourusername/opencursor.git
cd opencursor

# Install dependencies
poetry install

# Run tests
poetry run pytest

Project Structure

opencursor/
├── code_agent/
│   ├── src/
│   │   ├── agent.py      # Main agent implementation
│   │   ├── llm.py        # LLM client
│   │   ├── prompts.py    # System prompts
│   │   ├── tools.py      # Tool implementations
│   │   └── ...
│   ├── cli.py            # Command-line interface
│   └── __init__.py
├── tests/
├── pyproject.toml
└── README.md

License

MIT

UI Components

• Chat History: Shows the conversation between you and the AI
• Message Input: Type your messages here
• Tool Selection: Choose which tool to use for processing your message
• Workspace Path: Set the directory to work with
• Context Information: Shows which files are currently in context
• Update Context: Refreshes the context information
• Clear Chat: Clears the chat history

Available Tools

• agent (autonomous): Agent works step-by-step without user interaction
• agent (interactive): Agent performs one tool call at a time, waiting for user input
• chat (LLM only): Chat with the LLM directly without using tools
• add file: Add a file to the context (provide file path in message)
• drop file: Remove a file from the context (provide file path in message)
• clear context: Remove all files from the context
• repo map: Show the files in the current workspace
• focus on file: Add a file to context and show its contents

Customization

You can modify the model and host settings in the main() function of gradio_ui.py.