AgentBuilder 🤖

UI-first AI Agent Platform - Build, run, and observe AI agents with ease.

🌟 Overview

AgentBuilder is a modern, high-performance platform for creating and orchestrating AI agents. It provides a sleek UI to configure agents, define multi-agent workflows, and monitor execution with detailed logs and grounding metrics.

Built on the powerful agenticaiframework, AgentBuilder bridges the gap between complex agentic logic and user-friendly interfaces.

✨ Key Features

• 🤖 Visual Agent Builder: Configure roles, goals, and capabilities without writing code.
• 🔄 Workflow Orchestration: Design multi-agent teams (Sequential, Supervisor, Peer coordination).
• 📚 Advanced RAG (Knowledge Base):
  • Google Gemini Integration: Uses text-embedding-004 for high-fidelity retrieval.
  • Pinecone Vector Database: Scalable, high-performance vector search.
  • Strict Grounding: Built-in instructions to prevent hallucinations and cite sources.
• 🎮 Immersive Playground:
  • Prompt Studio: Sleek, glassmorphism-inspired interface for agent testing.
  • Grounded Results: View "Knowledge Atoms" and confidence scores for every RAG response.
• 📊 Observability: Real-time execution logs, step-by-step tool monitoring, and token/cost estimation.
• 🛠️ Extensible Tool Registry: Built-in tools for Web Search, Document Parsing, and more.

🚀 Quick Start

Backend (Python)

1. Install Dependencies:

   cd backend
   python -m venv venv
   source venv/bin/activate  # .\venv\Scripts\activate on Windows
   pip install -r requirements.txt

2. Configure Environment: Create a .env file in the backend/ directory:

   GEMINI_API_KEY=your_gemini_key
   PINECONE_API_KEY=your_pinecone_key
   PINECONE_INDEX_NAME=rag-index
   MONGO_URI=mongodb://localhost:27017
   MONGO_DB_NAME=agentbuilder

3. Run Server:

   uvicorn main:app --reload

Frontend (Next.js)

1. Install Dependencies:

   cd frontend
   npm install

2. Run Development Server:

   npm run dev

🏗️ Architecture

AgentBuilder uses a modern decoupled architecture:

• Core: agenticaiframework for agent/workflow execution.
• API: FastAPI providing a robust REST layer.
• Frontend: Next.js 14 with Tailwind CSS and Framer Motion for a premium DX.
• Vector Store: Pinecone for scalable RAG implementation.
• Database: MongoDB for agent configurations and execution history.

📖 Documentation

Detailed documentation is available in the /docs directory:

• System Architecture
• Tech Stack
• Application Flow
• API Reference
• RAG & Grounding (New!)

🤝 Contributing

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

📄 License

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