kubrick-cli 0.1.6

A CLI tool for AI-assisted coding using Triton LLM backend

Kubrick CLI

AI-assisted coding CLI tool supporting local LLMs via Triton, OpenAI, and Anthropic.

Installation

PyPI

pip install kubrick-cli

Docker

Run using the wrapper script (handles file permissions automatically):

curl -fsSL https://raw.githubusercontent.com/rcland12/kubrick-cli/master/scripts/install-kubrick-docker.sh | sh
kubrick-docker

Or run manually with docker run (include --user $(id -u):$(id -g) to prevent root-owned files):

docker run --rm -it \
  --network host \
  --user $(id -u):$(id -g) \
  -v ${HOME}:/home/kubrick \
  -v ${PWD}:/workspace \
  rcland12/kubrick-cli

See DOCKER.md for detailed setup and troubleshooting.

Quick Start

# Start Kubrick (runs setup wizard on first launch)
kubrick

# View available commands
You: /help

Example

You: Create a Python function to calculate fibonacci numbers with memoization

Assistant: I'll create a fibonacci function with memoization.

Tool: write_file
File: fibonacci.py

✓ Created fibonacci.py with memoized fibonacci function

Requirements

• Python 3.8+
• LLM Provider: Triton Inference Server (default), OpenAI API, or Anthropic API

Features

• Multi-provider support (Triton, OpenAI, Anthropic)
• Automatic file operations (read, write, edit)
• Code search with glob patterns and grep
• Bash command execution
• Real-time streaming responses
• Automatic context management
• Conversation history with persistence
• In-session configuration

Documentation

• WIKI.md - Complete guide, commands, and configuration
• PROVIDERS.md - Multi-provider setup
• DOCKER.md - Docker usage and troubleshooting
• TRITON.md - Triton backend setup
• CONTEXT_MANAGEMENT_QUICKSTART.md - Context window configuration
• TESTING.md - Development and testing

License

MIT License - See LICENSE