Changelog
Latest improvements, features, and fixes for Roopik.
Multi-Agent Support via MCP
Roopik now supports the Model Context Protocol (MCP), enabling external AI agents to connect and interact with the IDE. Run multiple agents from different providers simultaneously.
- NEWMCP Server: Built-in MCP server that connects Roopik IDE capabilities to external AI agents.
- NEWClaude Code and Codex Integration: Auto-registers with Claude Code & OpenAI Codex extension and CLI. Use Claude alongside Dio to build. Ultra Power Multiplier+
- NEWProject Mode: Built-in dev server with an embedded browser for running projects inside the IDE.
- NEWAgent skills: You can now write specific skills (custom instructions or code blocks) for Dio, and the agent will follow them to perform tasks or automate workflows. This makes Dio more flexible and lets you extend its capabilities for your own needs.
Initial Alpha Release
The first public preview of Roopik is now available for early access users.
Note: This alpha release has many known bugs. It's shared early to gather initial feedback on features which are actually relevant, beyond my imagination.
- NEWCanvas Mode: Infinite canvas to create and render multiple components efficiently in one space.
- NEWAgent Dio: AI coding assistant with deep workspace and browser awareness, powered by our own custom agent tools.
From Learning to Buidling Roopik
While exploring plugin-based software architectures, I discovered Code‑OSS and was immediately drawn to its flexibility and design. As I learned more, my focus shifted from building a personal assistant to creating an agentic IDE that could address the real challenges I encountered. That journey led to Roopik.
Personal AI Assistant & Distributed Agent Orchestration — Reached Top 10% (YC)
We were exploring distributed agent orchestration in the context of a personal AI assistant designed for individual users. The assistant could coordinate multiple specialized agents across different environments, with shared memory and modular skills enabling collaboration on tasks both within and beyond a single device ecosystem.