Welcome to Weave CLI

🚧 Experimental Project

Weave CLI is an experimental, AI-first CLI built to explore the principles of the Model Context Protocol.
It’s part of an ongoing learning journey — expect rapid changes, rough edges, and interesting ideas.

Why Weave?

Weave is a developer-first CLI designed to bring the power of structured context and AI to your local workflow.

Generate smarter tasks

Use GPT to generate high-level goals, tasks, and subtasks for any project idea or technical roadmap.

Break things down fast

Go from epic to detailed task breakdown with one command — Weave makes progress planning instant.

Get AI code help

Ask for help with any task and Weave will generate implementation advice and sample code using GPT.

Own your context

Everything is saved locally — context, tasks, history, and exports. No SaaS lock-in, just dev tools.

Export everything

Share project plans or task boards with Markdown or JSON export. Great for docs, clients, or your team.

---

What’s in the box?

• 🌱 weave init — start fresh with a local workspace
• 🧠 generate-tasks — generate epics using your project goals
• 🪓 breakdown — break down high-level tasks with GPT
• 💡 ai-help — get implementation help or boilerplate
• 📤 export — create sharable docs or project boards
• 🔐 .mcp/ + config — stores project and context state locally

---

📝 Example Output

Here’s a sample of what Weave CLI generates when exporting a project plan to Markdown:

Project: Real-time Chat App

Goals

• Build a secure, scalable real-time chat application
• Use Supabase for backend and Auth
• Deploy using Vercel

Epics & Tasks

🧠 Epic: Authentication

• Implement Supabase Auth with email/password
• Add OAuth via GitHub
• Handle JWT securely on client

🛠 Epic: Realtime Messaging

• Set up Supabase Realtime
• Create message schema and table
• Build WebSocket listener in frontend

🚀 Epic: Deployment

• Prepare vercel.json config
• Setup CI/CD with GitHub Actions
• Run final tests and ship

---

🤖 Example: ai-help

Let’s say you want help with this task:

Task: Build WebSocket listener in frontend

Reference the task using the id of task and subtasks

You can run:

weave ai-help 1.2

Weave will respond with something like:

🔧 Implementation Plan for: Build WebSocket listener in frontend

To implement a WebSocket listener in your frontend (React, Vue, or vanilla JS), follow these steps:

1. Create a WebSocket client using the WebSocket API:

   const socket = new WebSocket("wss://your-project.supabase.co/realtime/v1");

2. Handle open, message, and close events:

   socket.onopen = () => console.log("Connected to WebSocket");
   
   socket.onmessage = (event) => {
     const data = JSON.parse(event.data);
     // Handle incoming message
   };
   
   socket.onclose = () => console.log("Disconnected");

3. Authenticate if needed by appending your Supabase JWT as a query param or via headers (depending on config).

4. Reconnect on failure using an exponential backoff strategy or a library like reconnecting-websocket.

Every ai-help response is tailored to your task and local context — helping you move faster with less guesswork.