Built by Keshav Ramamurthy, Timothy Rolshud, Pragath Siva and Purnima Vasistha

Initial Public Offering System

Demo Media

• Video: Watch Video
• Presentation: View Presentation

To build ABIs for a smart contract: sforge build

To bring up the validator: sanvil

To deploy contracts and mint USDC for sample purposes, run: bun packages/cli/src/cli.ts setup

To create the new IPO Cross, run: bun packages/cli/src/cli.ts create

To get the current clearing price, run: bun packages/cli/src/cli.ts get-price <IPO Cross address>

To submit orders, run: bun packages/cli/src/cli.ts submit-orders <IPO Cross address>

To finalize the auction, run: bun packages/cli/src/cli.ts finalize <IPO Cross address>

---

Aptos Developer Chatbot

Demo Media

• Video: Watch Video
• Presentation: View Presentation

The Aptos Developer Chatbot is a multi-agent system designed to deliver precise, context-aware answers to Aptos developer queries. Leveraging retrieval-augmented generation (RAG) techniques, it extracts verified information directly from official Aptos resources—ensuring that answers are both accurate and relevant.

---

Overview

This solution consists of two key components:

• Backend: A Python/Flask server powered by the crewAI framework. It orchestrates a multi-agent workflow where specialized research agents query various Aptos-related data sources in parallel. Their results are combined into a single, cohesive markdown-formatted response. The backend also manages API endpoints and conversation state, allowing for both single-turn and multi-turn interactions.

• Frontend: A modern, TypeScript-based chatbot UI built with Next.js. The interface provides features such as suggested questions, conversation history, and an easily embeddable design for integration into existing Aptos documentation platforms like aptos.dev.

---

Multi-Agent Workflow & Architecture

The heart of the system is its multi-agent architecture, which operates as follows:

1. API Interaction & Context Management:

   • RESTful endpoints (e.g., /ask, /conversation) handle incoming queries and maintain conversation history.
   • Session context is preserved across interactions to provide coherent multi-turn dialogues.

2. Parallel Research with Specialized Agents:

   • The backend instantiates several research agents (each configured via YAML) to search distinct Aptos resources, such as core repositories, developer docs, and tutorials.

   • Using asyncio, these agents run concurrently, each performing its assigned research task to gather targeted information.

   • The concurrent operation ensures minimal latency and broad coverage of different information sources.

     Sources:

     • 'https://github.com/aptos-labs/aptos-core'
     • 'https://github.com/0xaptosj/aptos-full-stack-template'
     • 'https://github.com/aptos-labs/developer-docs'
     • 'https://learn.aptoslabs.com/'
     • 'https://aptos.dev/'

3. Results Aggregation & Synthesis:

   • A dedicated synthesis process (reporting workflow) aggregates outputs from all agents.
   • This process uses retrieval-augmented generation (RAG) techniques to combine and clean the data into a final, concise answer formatted in markdown.
   • The final response integrates essential insights from each agent, ensuring the answer is comprehensive and developer-friendly.

---

Frontend Integration

• Chatbot UI Features:
  Built with Next.js and TypeScript, the UI offers an intuitive chat interface, featuring:

  • Suggested Questions: Predefined queries (e.g., “How to deploy a Move module?”) to prompt user interaction.
  • Conversation History: Display of previous interactions to maintain context.
  • Embed-Ready Design: Easily integrated into existing Aptos documentation platforms for a seamless developer experience.

• Seamless Communication:
  The frontend communicates with the backend via REST API endpoints, ensuring prompt and interactive query responses.

---

Setup and Installation

Backend Setup

1. Prerequisites:

   • Python (>=3.10, <3.13)
   • Required Python packages installed via pip
   • Environment variables configured in a .env file (e.g., OPENAI_API_KEY, GitHub tokens)

2. Installation:

   pip install -r requirements.txt
   

3. Run the Backend:

   python backend/aptos/src/aptos/main.py

   The API will be available at http://localhost:8007.

Frontend Setup

1. Prerequisites:

   • Node.js and npm (or yarn)

2. Installation:

   cd aptos-ui
   npm install

3. Run the Frontend:

   npm run dev

   Access the chatbot UI at http://localhost:3000.

---

Conclusion

The Aptos Developer Chatbot represents a robust, multi-agent architecture that efficiently answers technical questions by aggregating data from trusted Aptos resources. Its parallel research workflow minimizes response time while ensuring accuracy, and its TypeScript-based UI offers a seamless, feature-rich user experience. This makes the chatbot an ideal tool for integrating into existing Aptos developer portals, enhancing support and productivity.

Happy coding, and enjoy your enhanced Aptos experience!