AI2Apps Agents

AI2Apps Agents is a collection of AI agents and tools that support speech processing, text-to-speech, chemistry research, image generation, web automation, and more.

All the agents are built in AI2Apps platform.

📁 Project Structure

Core Components

• AgentBuilder - Agent construction tool providing project setup, environment configuration, code generation, and more
• gernalist.js - General-purpose agent supporting tool management and speed mode
• pkgUtil.js - Package management utility
• setup.js - Project setup script

AI Agents

Speech Processing

• whisper - Speech recognition agent based on OpenAI Whisper model
• coqui_tts - Text-to-speech agent using Coqui TTS
• mega_tts3 - Advanced text-to-speech service
• sparktts - iFLYTEK speech synthesis service

Professional Domains

• ChemAgent - Chemistry research agent with speed mode and expert mode
• chemformer - Chemical molecule transformation tools
• chemt5 - Chemical text processing model

Image & Vision

• DrawSD - Image generation agent based on Stable Diffusion
• sam2 - Image segmentation tools

Web & Automation

• WebAgent - Web RPA agent supporting web automation
• WebRpa - Web robotic process automation
• DeepSearch - Deep search agent

Communication & Integration

• wechat - WeChat integration agent

Deployment & Testing

• AutoDeploy - Automated deployment tools
• SetupTest - Setup testing tools

🚀 Quick Start

1. Follow the guide in AI2Apps to deploy the AI2Apps platform locally

2. Create an empty backend agent within the AI2Apps platform environment and open it in an IDE

3. Sync/copy the contents of the specific agent you want to use to the ai/ subdirectory

🔧 Main Features

General Agent (gernalist)

• Tool management and loading
• Speed mode operations
• Multi-language support (English/Chinese)
• Dynamic tool selection and execution

Chemistry Research Agent (ChemAgent)

• Chemical molecule analysis
• Reaction prediction
• Literature search
• Experimental design suggestions

Speech Processing Agents

• Real-time speech recognition
• Multi-language TTS
• Audio format conversion
• Speech quality optimization

Image Generation Agent (DrawSD)

• Text-based image generation
• Image editing and modification
• Multiple artistic style support
• Batch image processing

Web Automation Agent (WebAgent)

• Web data scraping
• Automated operations
• Form filling
• Content monitoring

🌐 Configuration

agenthub.json Configuration

{
    "user": "your-email@example.com",
    "devKey": "YOUR-Dev-Key",
    "host": "ws://localhost:3015",
    "language": "EN",
    "globalContext": {
        "models": {
            "expert": {"platform": "OpenAI", "model": "gpt-4o"},
            "standard": {"platform": "OpenAI", "model": "gpt-4o"},
            "lite": {"platform": "OpenAI", "model": "gpt-4o-mini"}
        }
    }
}

🔒 Privacy & Security

This repository has been carefully cleaned to ensure no sensitive information is exposed:

• Sensitive directories excluded: Certain directories containing sensitive code (like WeChat integration details) are excluded from version control
• Git history cleaned: All sensitive information has been completely removed from Git history
• Privacy-first approach: Only safe, non-sensitive code is publicly available

Excluded Components

• Sensitive integration code and configurations
• API keys and authentication tokens
• Private user data and credentials

🔄 Development Mode

Project Setup

Use setup.js for project configuration:

import {setupPrj, setupAgent} from './setup.js';

// Setup project
await setupPrj(session, path, "EN");

// Setup agent
await setupAgent(session, path, "EN");

Adding New Agents

1. Create agent files in appropriate directory
2. Configure agent.json
3. Implement necessary interface functions
4. Update main configuration

📝 Usage Instructions

1. Select Agent: Choose appropriate agent based on requirements
2. Configure Environment: Ensure all dependencies are installed
3. Set API Keys: Configure corresponding API access permissions
4. Run Agent: Start the corresponding agent service
5. Interactive Use: Interact with agents through chat interface or API

---

Last updated: 2025-01-09