Not an AI assistant. An AI engineering fleet.

---

⚡ Parallel Agents • 🧠 AI-Native • 🔧 Developer First • 🚀 Production Ready

---

✨ What is ForgeFlow?

ForgeFlow is a multi-agent AI engineering system that:

• 🧩 Debugs production issues from logs
• 🏗️ Generates full-stack app blueprints
• ⚡ Runs 5 parallel AI agents
• 🔌 Uses an MCP server for intelligent tooling

---

🧠 The One-Line Pitch

ForgeFlow Fleet turns a messy error log, blurry screenshot, or vague idea into a complete, production-ready Vercel AI agent blueprint — debugged, designed, and ready to ship — in under 60 seconds.

---

🎯 Problem → Solution

Developers face two daily pains:

1. Production debugging hell — Cryptic logs and errors that waste hours.
2. Idea-to-implementation gap — Struggling to structure AI agents quickly on the Vercel stack.

ForgeFlow deploys a 5-specialist AI fleet that works in parallel using Vercel AI SDK, delivering results in real time.

📊 5 Specialized AI Agents

Agent	Expertise	Output
🔍 Analyzer	Root cause analysis, idea refinement	Incident summary, severity, technical context
🎨 UI Architect	Interface design, v0 prompts	2-3 production-ready v0.dev prompts
⚙️ Backend Engineer	System architecture, MCP tooling	Architecture blueprint, 3-5 custom MCP tools
⚖️ Judge	Solution evaluation, quality scoring	Usefulness, Technical Execution, Creativity scores (1-10)
🎬 Scriptwriter	Demo creation, storytelling	45-60 second pitch-perfect demo script

---

🛠️ Tech Stack

Layer	Technology	Purpose
Frontend	Next.js 15 (App Router) + Tailwind CSS	Streaming‑ready, edge‑optimized dashboard
UI Components	React (Client Components) + Framer Motion	Smooth state management (Profile, Settings, Autonomous Mode)
AI Orchestration	Vercel AI SDK (streamObject) + Server Actions	5 parallel agent streams bypassing API route overhead
AI Models	Anthropic Claude 3.5 Sonnet	Deep reasoning, dynamic prompt rewriting (/evolve)
Persistence	Vercel KV	Incident history, prompt versioning (self‑evolution)
Deployment	Vercel	Instant global edge deployment

🧪 Live Demo

---

✨ Key Features

🚀 Parallel AI Execution

Five specialized agents run simultaneously using Promise.all(), not sequentially. See results stream in real-time as each agent completes its analysis.

🎨 Built Entirely on Track 2 Tech

• v0.dev: UI components designed and iterated with v0 prompts
• Vercel AI SDK: Structured outputs with generateText and Zod schemas
• Custom MCP Server: 4+ production-ready tools (log parser, error analyzer, fix generator, v0 prompt suggester)

🖼️ Multimodal Intelligence

• Paste error logs from production
• Upload screenshots of broken UIs (vision analysis)
• Type natural language ideas
• Voice input support

💎 Why This Wins Hackathons

Criterion	Proof
Usefulness	Solves real pains for every developer (debugging + rapid AI prototyping)
Technical Execution	Next.js 15 + Vercel AI SDK + custom MCP + Vercel KV + parallel streaming
Creativity	First visual parallel AI fleet with self-healing simulation and self-evolution

---

🏗️ Architecture

flowchart TD
    A["User Input<br/>Text, Logs, Screenshots, Voice"] --> B["/api/forgeflow"]

    B --> C1["runAnalyzer()"]
    B --> C2["runUIArchitect()"]
    B --> C3["runBackendEng()"]
    B --> C4["runJudge()"]
    B --> C5["runScriptwriter()"]

    C1 --> D["MCP Server (/api/mcp)"]
    C2 --> D
    C3 --> D
    C4 --> D
    C5 --> D

    D --> E["parse_log"]
    D --> F["analyze_error"]
    D --> G["generate_fix"]
    D --> H["suggest_v0_prompt"]

Loading

---

## ⚙️ How It Works (Under the Hood)
1. User input enters the **Input Card** and checks for special commands (like `/evolve`)
2. **Server Action** (`launchFleetAction`) invokes the Vercel AI SDK with your provided Anthropic API key
3. The Action orchestrates 5 parallel `streamObject` calls, generating typed JSON structures
4. The client uses `readStreamableValue` to consume these streams simultaneously
5. The UI dynamically toggles between "Generating..." skeletons and the rich, typed `AgentPreview` UI
6. If **Autonomous Mode** is enabled, a `useEffect` loop automatically waits and re-launches the fleet with the next evolutionary step

---

🚀 Quick Start

git clone https://github.com/AbhishekGupta0164/ForgeFlow-Multi-Agentic-AI-System
cd ForgeFlow-Multi-Agentic-AI-System

npm install

cp .env.local.example .env.local
# Add your ANTHROPIC_API_KEY

npm run dev

👉 Open: http://localhost:3000

---

📊 Project Structure

textforgeflow/
├── app/
│   ├── layout.tsx         # Root layout (fonts + metadata)
│   ├── page.tsx           # Main dashboard (input + agents + blueprint)
│   ├── globals.css        # Theme, animations, utilities
│   └── api/
│       ├── forgeflow/
│       │   └── route.ts   # Fleet orchestration (5 parallel agents)
│       └── mcp/
│           └── route.ts   # MCP server (4 tools)
│
├── components/
│   ├── AgentCard.tsx      # Animated agent status cards
│   └── BlueprintPanel.tsx # Tabbed results viewer
│
├── lib/
│   ├── schemas.ts         # Zod schemas for agents
│   └── prompts.ts         # System prompts
│
├── package.json
├── next.config.js
├── tailwind.config.ts     # Custom theme
└── tsconfig.json

---

🧪 Example Outputs

🧩 Debugging Case

Input: Redis connection refused error log

Output:

• Root cause analysis
• Monitoring dashboard (v0 prompt)
• Retry logic + MCP tool
• Judge scores
• Demo script

---

🤖 AI Builder Case

Input:
Build a commit message AI agent from git diffs

Output:

• Full system blueprint
• VS Code-style UI
• Parser tools
• Agent pipeline

---

📋 Roadmap

• ⚡ Parallel agents + streaming
• 🧠 Multimodal + MCP integration
• 🎤 Voice input
• 🔗 GitHub export
• 👀 Live preview
• 🤖 Self-evolving agents

---

🤝 Contributing

Contributions are welcome!
See CONTRIBUTING.md

---

📝 License

MIT License

---

❤️ Built For

v0 → Agent Hackathon

Transform chaos into clarity.
Turn ideas into shippable AI agents.

---