Creo

Creo is a native desktop IDE purpose-built for developing Chainlink CRE
 workflows — with a focus on AI-in-the-loop applications and autonomous Web3 agents.          
                                                    
 The problem: Building CRE workflows today means writing Go code blind — no tooling, no live feedback,
 no way to test trigger logic without deploying to the DON, and no bridge between AI systems and
 on-chain execution. Developers waste hours on boilerplate, misconfigured EVM triggers, and
 trial-and-error simulation cycles.

 How Creo solves it: Creo brings everything into one environment. You write your workflow in a
 Monaco-powered Go editor, configure EVM triggers by pasting an ABI and selecting events, simulate
 against real chain conditions locally, and deploy — all without leaving the app. An embedded AI agent
 (Agent Architect) reads your current workflow files, understands your trigger config, and edits your
 code directly from natural language. The x402 tab lets you simulate exactly how an AI agent would pay
 per-call using EIP-3009 permit signatures before triggering a workflow — closing the loop between AI
 decision-making and verifiable on-chain execution

Creo is built as a Wails v2 desktop application embedding a Next.js 15 frontend. The Go backend
 handles Docker-based workflow compilation (targeting wasip1/wasm for CRE compatibility), host-side
 simulation via the cre CLI, and EVM binding generation. The frontend is a full IDE shell: Monaco
 editor, a resizable panel layout, an EVM trigger builder with ABI parsing, a real-time simulation
 stepper, and an interactive shell terminal. The Agent Architect uses OpenRouter (with a
 user-selectable model from GPT-4o to Claude to DeepSeek) backed by Pinecone vector search over CRE
 documentation for grounded, RAG-augmented code edits. File edits from the AI agent are applied
 directly to the in-memory file tree and reflected instantly in the editor.