PiClaude

PiClaude is an open-source coding-agent CLI for cloud and local model providers. It allows you to run, build, and edit code via automated terminal workflows using any LLM backend.

Features

• Multi-Provider Support — Run with any provider of your choice, including OpenAI-compatible APIs, Gemini, Ollama, and local models.
• Terminal-First Workflow — Automated coding tools including bash execution, file read/write/edit, grep search, and web search/fetch.
• Slash Commands — Manage settings, models, connection status, and tasks directly from the interactive chat interface.

Table of Contents

• Features
• Table of Contents
• Installation
• Usage
  • Start
  • Configuration

Installation

Install the CLI globally via npm:

npm install -g @rynn-k/piclaude@latest

Usage

Start

Launch the interactive CLI by running:

piclaude

Type /help inside the session to browse all available commands (such as /provider, /model, /config, /diff, /tasks, and /exit).

Configuration

You can set up your backend interactively or by using environment variables.

Interactive Setup: Run /provider inside PiClaude to choose and save your API endpoint, key, and model.

Environment Variable Setup: Bypass setup by exporting your endpoint credentials:

export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_BASE_URL="https://api.your-anthropic.com/v1"
export OPENAI_API_KEY="your-api-key"
export OPENAI_MODEL="your-model-name"

Disclaimer: PiClaude is an independent community project and is not affiliated with Provider.

License: MIT