Skip to content

Repository files navigation

🤖 AgentForge

Multi-Agent AI Planning Engine

Python FastAPI React LangGraph Docker License

Give it a business idea. Watch 7 AI agents debate, analyze, and build your startup blueprint — live.

Features · Architecture · Quick Start · Tech Stack · Demo


🎯 What is AgentForge?

AgentForge is a production-grade multi-agent AI system that transforms a raw business idea into a comprehensive startup blueprint. It orchestrates 7 specialized AI agents through a sophisticated LangGraph state-machine workflow, featuring parallel execution, multi-step reasoning, and a unique Board Meeting debate mechanism for cross-validation.

💡 Example: You type "AI-powered fitness coaching app" → AgentForge deploys 7 agents simultaneously → Within minutes, you get a 30-page PDF blueprint covering market research, competitive analysis, technical architecture, financial projections, go-to-market strategy, operations plan, and risk assessment.


✨ Features

Feature Description
🧠 7 Autonomous Agents CEO, Research, Marketing, Developer, Finance, Analytics, Operations — each with domain-specific prompts and tools
💎 Diamond Topology Fan-out/fan-in parallel execution via LangGraph state machine (4 agents run simultaneously)
🔄 Multi-Step Reasoning Each agent runs a 3-step pipeline: Generate → Self-Critique → Refine with confidence scoring
Real-Time Streaming WebSocket-powered live visibility into agent thinking, reasoning steps, and inter-agent messages
🏛️ Board Meeting Debate Unique 4-round cross-functional debate where agents challenge each other's assumptions
📄 Auto-Generated Reports Beautiful PDF blueprints compiled from all agent outputs with charts and data
🔐 JWT Authentication Secure user sessions with protected routes and role-based access
🐳 Docker Compose One-command deployment with PostgreSQL, Redis, and all services containerized

🏗 Architecture

┌─────────────────────────────────────────────────────────────┐
│                   PRESENTATION TIER                         │
│              React + Vite + WebSocket Client                │
│   ┌──────────┐  ┌──────────────┐  ┌──────────────┐         │
│   │Dashboard │  │ Live Stream  │  │Report Viewer │         │
│   └──────────┘  └──────────────┘  └──────────────┘         │
└─────────────────────────┬───────────────────────────────────┘
                          │ REST + WebSocket
┌─────────────────────────┴───────────────────────────────────┐
│                   APPLICATION TIER                          │
│              FastAPI + LangGraph + Claude/Gemini            │
│                                                             │
│   ┌─────────┐    ┌───────────────┐    ┌──────────────┐     │
│   │REST API │───▶│  Orchestrator │───▶│  AI Service   │     │
│   └─────────┘    └───────┬───────┘    └──────────────┘     │
│                          │                                   │
│              ┌───────────┼───────────┐                      │
│              ▼           ▼           ▼                      │
│         ┌────────┐ ┌────────┐ ┌────────┐                   │
│         │Research│ │  Dev   │ │Finance │  ... (7 agents)    │
│         └────────┘ └────────┘ └────────┘                   │
│                          │                                   │
│              ┌───────────┴───────────┐                      │
│              ▼                       ▼                      │
│        ┌──────────┐          ┌────────────┐                 │
│        │Board Meet│          │Report Gen  │                 │
│        └──────────┘          └────────────┘                 │
└─────────────────────────┬───────────────────────────────────┘
                          │
┌─────────────────────────┴───────────────────────────────────┐
│                      DATA TIER                              │
│           PostgreSQL + Redis + SQLAlchemy ORM               │
└─────────────────────────────────────────────────────────────┘

🔀 Agent Pipeline (9-Node Diamond Topology)

START → CEO Agent → ┬─ Research Agent  ─┐
                    ├─ Marketing Agent ─┤
                    ├─ Developer Agent ─┤─→ Analytics → Operations → Board Meeting → Report → END
                    └─ Finance Agent   ─┘
                      (parallel fan-out)    (fan-in)        (sequential)         (debate)

🧠 Multi-Step Reasoning (Per Agent)

┌──────────────┐     ┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│  Step 1:     │────▶│  Step 2:     │────▶│  Step 3:     │────▶│  Validate:   │
│  Generate    │     │  Self-Critique│     │  Refine      │     │  Pydantic    │
│  (Claude API)│     │  (3 flaws)   │     │  (Fix flaws) │     │  (JSON safe) │
└──────────────┘     └──────────────┘     └──────────────┘     └──────────────┘
       │                    │                    │
   🔴 stream:           🔴 stream:           🔴 stream:
   agent_thinking       reasoning_step       agent_completed

🔄 User Flow Diagram

sequenceDiagram
    autonumber
    actor User as Startup Founder
    participant UI as React UI (WebSocket Client)
    participant API as FastAPI Gateway
    participant LG as LangGraph Orchestrator
    participant AG as 7-Agent Cluster
    participant DB as PostgreSQL Database

    User->>UI: Submits raw startup idea
    UI->>API: POST /api/blueprint (Initialize Session)
    API->>DB: Create project record & audit trail
    API->>LG: Trigger graph orchestration (background task)
    LG->>UI: Establish WebSocket connection
    
    rect rgb(30, 41, 59)
        note over LG, AG: Parallel Agent Execution (Diamond Topology)
        LG->>AG: Fan-out: Invoke CEO, Research, Marketing, Dev, & Finance agents
        AG-->>UI: Live-stream thinking tokens & reasoning steps via WS
    end

    rect rgb(15, 23, 42)
        note over LG, AG: Consensus Debate & Report Generation
        LG->>AG: Initiate structured 4-Round Board Meeting Debate
        AG-->>UI: Broadcast inter-agent debate statements
        LG->>API: Compile final consolidated markdown report
        API->>DB: Save generated PDF/Markdown blueprint files
    end

    API-->>UI: Emit workflow completion signal
    UI->>User: Render 3D debate visualization & enable PDF download
Loading

🚀 Quick Start

Prerequisites

  • Python 3.11+
  • Node.js 18+
  • Docker & Docker Compose
  • Claude API key or Gemini API key

1. Clone & Configure

git clone https://github.com/rajmodi262/agentforge.git
cd agentforge
cp .env.example .env
# Edit .env with your API keys

2. Run with Docker (Recommended)

docker-compose up --build

3. Run Locally (Development)

# Backend
cd backend
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

# Frontend (new terminal)
cd frontend
npm install
npm run dev

4. Access

Service URL
🌐 Frontend http://localhost:5173
⚡ Backend API http://localhost:8000
📚 API Docs http://localhost:8000/docs

🛠 Tech Stack

Layer Technology Purpose
🧠 AI Engine LangGraph, Claude API, Gemini API Agent orchestration, multi-step LLM reasoning
⚡ Backend FastAPI, SQLAlchemy, Pydantic Async REST API, ORM, schema validation
🔄 Real-Time WebSocket, Redis Live agent streaming, event broadcasting
🎨 Frontend React 18, Vite, GSAP Interactive dashboard, 3D animations
🗄️ Database PostgreSQL, SQLite (dev) Persistent storage, session management
🐳 DevOps Docker Compose, GitHub Actions Containerized deployment, CI/CD

📂 Project Structure

agentforge/
├── backend/
│   ├── app/
│   │   ├── agents/              # 7 AI agents + orchestrator
│   │   │   ├── base_agent.py    # Multi-step reasoning engine
│   │   │   ├── orchestrator.py  # LangGraph workflow builder
│   │   │   ├── board_meeting.py # Multi-round debate engine
│   │   │   ├── ceo_agent.py
│   │   │   ├── research_agent.py
│   │   │   ├── marketing_agent.py
│   │   │   ├── developer_agent.py
│   │   │   ├── finance_agent.py
│   │   │   ├── analytics_agent.py
│   │   │   ├── operations_agent.py
│   │   │   └── report_compiler.py
│   │   ├── api/                 # REST endpoints
│   │   ├── models/              # Pydantic schemas + DB models
│   │   ├── services/            # Claude/Gemini, WebSocket, RAG
│   │   └── main.py              # FastAPI application entry
│   ├── tests/                   # Pytest test suite
│   └── requirements.txt
├── frontend/                    # React + Vite SPA
├── docker-compose.yml
└── README.md

🏛️ Board Meeting — The Secret Sauce

What makes AgentForge unique is the Board Meeting Debate Engine. After all 7 agents complete their analysis, they enter a structured 4-round debate:

Round Matchup Topic
1 Finance → CEO Revenue model & unit economics viability
2 Developer → Marketing Go-to-market timeline vs. technical feasibility
3 Research → Finance TAM/SAM accuracy & market sizing assumptions
4 Operations → Developer Infrastructure scaling readiness

A Board Secretary then synthesizes all debates into a consensus score, identifies key risks, and produces the final strategic recommendation.


📊 What You Get

When you submit a business idea, AgentForge generates a comprehensive blueprint including:

  • Vision & Problem Statement — Clear articulation of the opportunity
  • Market Research — TAM/SAM/SOM analysis with competitor mapping
  • Go-to-Market Strategy — Pricing, channels, customer acquisition cost
  • Technical Architecture — Full stack recommendation with MVP roadmap
  • Financial Projections — 3-year revenue model with break-even analysis
  • KPI Dashboard — Key metrics and success criteria
  • Operations Plan — Hiring roadmap, launch timeline, risk mitigation
  • Board Consensus — Cross-validated insights from agent debates

🔧 Environment Variables

# AI Provider (at least one required)
ANTHROPIC_API_KEY=sk-ant-...       # Claude API
GOOGLE_API_KEY=AIza...             # Gemini API (free tier)

# Database
DATABASE_URL=postgresql://user:pass@localhost/agentforge

# Security
JWT_SECRET_KEY=your-secret-key

# Optional
MOCK_MODE=true                     # Use dynamic mocks (no API cost)
LOG_LEVEL=INFO

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add amazing feature')
  4. Push to the branch (git push origin feature/amazing-feature)
  5. Open a Pull Request

📄 License

This project is licensed under the MIT License — see the LICENSE file for details.


Built with ❤️ by Raj Modi

If you found this useful, please ⭐ the repo!

About

Multi-Agent AI Planning Engine — 7 autonomous LLM agents orchestrated via LangGraph

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages