Inspiration

The inspiration for AIdeaForge came from observing how traditional brainstorming sessions often struggle with limited perspectives, individual biases, and difficulty connecting disparate ideas. We aimed to overcome these limitations by leveraging multiple specialized AI agents working collaboratively—akin to how diverse teams with varied expertise often yield superior solutions compared to individuals working alone. Microsoft's AutoGen framework provided the ideal foundation for this multi-agent collaboration system.

What it does

AIdeaForge is an AI-powered collaborative brainstorming platform where five specialized AI agents—Creative, Technical, Business, Critic, and Synthesizer—work together to generate innovative ideas on any topic. Users input a topic, industry, and optional constraints, then watch in real-time as agents collaborate, generating ideas, evaluating them from multiple perspectives, identifying potential issues, and refining them into polished concepts. The platform visualizes this collaboration through an interactive 3D graph and presents refined ideas alongside comprehensive evaluations. Users can rate ideas and provide feedback to further guide the process.

How we built it

We constructed AIdeaForge using cutting-edge AI and web technologies:

Backend

  • Microsoft's AutoGen framework for multi-agent orchestration
  • Custom prompt engineering for each specialized agent
  • Express.js server with Socket.IO for real-time communication
  • Idea extraction and processing algorithms

Frontend

  • Next.js for responsive and performant UI
  • Shadcn UI components with custom animations
  • Tailwind CSS for styling
  • Framer Motion for fluid animations
  • React Force Graph for 3D network visualization
  • Socket.IO client for real-time updates

The system extracts structured ideas, evaluations, and refinements from agent conversations, mapping these relationships into an interactive visual network that users can explore.

Challenges we ran into

  • Orchestrating multi-agent interactions: Ensuring productive collaboration required meticulous prompt design and structured conversation flow, balancing creative freedom with organized brainstorming.
  • Real-time visualization: Converting text-based agent conversations into meaningful visual representations involved complex graph generation algorithms and extensive performance optimization for smooth 3D rendering.
  • Structured data extraction: Developing effective pattern recognition systems to identify and structure new ideas, evaluations, and refinements from free-flowing agent discussions.

Accomplishments we're proud of

We successfully developed a collaborative AI system in which specialized agents effectively generate and refine ideas, mimicking human brainstorming but enhancing capabilities significantly. The 3D visualization vividly illustrates collaboration, enhancing both understanding and visual appeal.

We take particular pride in how the diverse agent personalities complement each other, creating a robust pipeline—from innovative concept generation by the Creative Agent to practical assessments by the Technical and Business Agents, critical evaluations by the Critic Agent, and refined synthesis by the Synthesizer Agent.

What we learned

We gained valuable insights into multi-agent AI system design, particularly in creating effective collaborative workflows and advanced prompt engineering techniques for specialized agent behaviors. Additionally, we honed technical skills in real-time data visualization, Socket.IO communication, and integrating AI systems within modern web frameworks.

What's next for AIdeaForge

Future enhancements planned for AIdeaForge include:

  • Custom agent creation, allowing users to define their own specialized agents.
  • Integration with project management tools to transform ideas into actionable plans.
  • Collaborative sessions, enabling multiple human users to brainstorm alongside AI agents.
  • Sophisticated idea evaluation metrics and scoring systems.
  • Industry-specific templates and knowledge bases to enrich domain expertise.
  • Export functionality for detailed idea documentation.

Built With

  • JavaScript/TypeScript
  • React
  • Next.js
  • Node.js
  • Express
  • Socket.IO
  • Microsoft AutoGen
  • OpenAI API
  • Shadcn UI
  • Tailwind CSS
  • Framer Motion
  • Three.js
  • React Force Graph

Built With

Share this project:

Updates