██████╗ ██╗   ██╗ ██████╗ ██████╗ ██████╗ ███████╗
██╔══██╗╚██╗ ██╔╝██╔════╝██╔═══██╗██╔══██╗██╔════╝
██████╔╝ ╚████╔╝ ██║     ██║   ██║██║  ██║█████╗
██╔══██╗  ╚██╔╝  ██║     ██║   ██║██║  ██║██╔══╝
██║  ██║   ██║   ╚██████╗╚██████╔╝██████╔╝███████╗
╚═╝  ╚═╝   ╚═╝    ╚═════╝ ╚═════╝ ╚═════╝ ╚══════╝

The AI-Native Evolution of OpenCode
Multi-agent development powered by toolkit-cli

---

🤖 The AI-Native Difference

$ toolkit-cli fix opencode/ --ai "claude gemini qwen codex"

● Claude analyzing architecture patterns...
● Gemini reviewing UX implications...
● Qwen checking i18n compliance...
● Codex generating optimized solution...

✓ Synthesized fix applied. 4 agents collaborated.

---

---

About RyCode

RyCode is the AI-native version of OpenCode - rebuilt from the ground up with multi-agent collaboration at its core.

While OpenCode pioneered AI-powered coding assistants, RyCode takes it further by making every aspect of development AI-native. Built using toolkit-cli's multi-agent architecture, RyCode doesn't just use AI - it's designed by AI collaboration.

Instead of a single AI perspective, Claude, Gemini, Codex, and Qwen work together - each bringing specialized expertise to architecture, performance, security, and optimization.

The Evolution Story

🔷 OpenCode - The Foundation A powerful terminal-based AI coding assistant, built with a single AI perspective.

✨ RyCode - The AI-Native Evolution OpenCode's codebase transformed through toolkit-cli's multi-agent collaboration:

/fix      # Claude identifies and fixes architectural issues
/improve  # Gemini suggests performance optimizations
/security # Qwen hardens security vulnerabilities
/optimize # Codex refactors for efficiency

What toolkit-cli Commands Were Run

RyCode showcases these toolkit-cli capabilities:

• /fix - Multi-agent bug detection and resolution
• /improve - Collaborative code quality enhancement
• /security - AI-powered security hardening
• /optimize - Performance and bundle size optimization
• /peer-review - Multi-perspective code review
• /test - Comprehensive testing strategy

Why Multi-Agent Optimization Works

Instead of one AI making all decisions, toolkit-cli orchestrates specialists:

• Claude → Deep reasoning and architecture analysis
• Gemini → Fast iteration and performance insights
• Codex → Precise code generation and refactoring
• Qwen → Efficient optimization and security

Result: Better code than any single AI could produce alone.

---

Installation

# Package managers
npm i -g rycode@latest             # or bun/pnpm/yarn
brew install aaronmrosenthal/tap/rycode      # macOS and Linux

Tip

RyCode is actively developed. Check back for updates!

Installation Directory

The install script respects the following priority order for the installation path:

1. $RYCODE_INSTALL_DIR - Custom installation directory
2. $XDG_BIN_DIR - XDG Base Directory Specification compliant path
3. $HOME/bin - Standard user binary directory (if exists or can be created)
4. $HOME/.rycode/bin - Default fallback

Documentation

For more info on how to configure RyCode, see the inline help and configuration options.

Contributing

RyCode is built with toolkit-cli and showcases multi-agent AI development. Contributions are welcome!

We accept PRs for:

• Bug fixes
• Improvements to LLM performance
• Support for new AI providers
• Env-specific fixes
• Documentation improvements
• New toolkit-cli slash command integrations

To run RyCode locally you need:

• Bun
• Golang 1.24.x

And run:

$ bun install
$ bun dev

Development Notes

Built with toolkit-cli: This project demonstrates collaborative AI development patterns. Changes to core functionality may involve multiple AI agents working together through toolkit's slash command system.

FAQ

What is RyCode?

RyCode is the AI-native version of OpenCode - rebuilt from the ground up using toolkit-cli's multi-agent collaboration.

It's not just optimized code; it's a complete evolution that demonstrates what happens when you design software with AI-first principles, where multiple AI agents collaborate on architecture, security, performance, and code quality from day one.

How was RyCode created?

# Started with OpenCode
git clone https://github.com/aaronmrosenthal/rycode

# Ran toolkit-cli optimization commands
toolkit-cli /fix opencode/
toolkit-cli /improve opencode/
toolkit-cli /security opencode/
toolkit-cli /optimize opencode/

# Result: RyCode - The AI-native version

What's different from OpenCode?

• AI-Native Architecture - Designed from the ground up with multi-agent collaboration
• Multi-Agent Foundation - Every feature built with Claude, Gemini, Codex, and Qwen working together
• Collective Intelligence - Not just one AI's perspective, but 4 specialized agents collaborating
• Systematically Evolved - 26 toolkit-cli commands applied across the entire codebase
• Living Demonstration - Showcases what's possible when AI agents collaborate on software design

Why multi-agent over single-AI?

Single AI: One perspective, one set of biases, one approach.

Multi-agent (toolkit-cli):

• Claude catches architectural issues
• Gemini spots performance bottlenecks
• Codex suggests better patterns
• Qwen finds security vulnerabilities

Result: Code that's been reviewed and optimized from 4 different expert perspectives.

---

Built with toolkit-cli | Created by Aaron Rosenthal