DATA-AI CLI

A Production-Ready CLI AI Agent Framework with Advanced Capabilities

English | 中文

---

🚀 Quick Start

# Install from PyPI
pip install data-ai

# Or install from source
git clone https://github.com/badhope/DATA-AI.git
cd DATA-AI
pip install -e ".[all]"

# Run the CLI
data-ai

---

✨ Core Features

🤖 Multi-Model Support

• Global Models: OpenAI GPT-4, Anthropic Claude, Google Gemini, Azure OpenAI, AWS Bedrock
• Chinese Models: DeepSeek, Qwen, Zhipu AI, Baichuan, Moonshot, SiliconFlow
• Local Models: Ollama, vLLM, LM Studio

🧠 Advanced Agent System

Component	Description
Intent Recognition	24+ intent types with intelligent routing
Sequential Thinking	Multi-step reasoning with hypothesis generation
Skill System	Auto-extraction, progressive disclosure (L0/L1/L2)
Code Knowledge Graph	Multi-language parsing, dependency tracking
Closed-Loop Learning	Execute → Evaluate → Extract → Retrieve

🛠️ Tool Ecosystem (60+ Tools)

Category	Tools
File Operations	Read, write, search, replace, batch operations
Code Execution	Python sandbox, Bash, Node.js
Browser	Navigation, click, screenshot, form filling
Search	DuckDuckGo, multi-engine aggregation
Documents	PDF, Word, Excel, PPT, Markdown, HTML
Data Analysis	Pandas, Matplotlib, Plotly

📚 RAG System

• Document Loaders: PDF, Word, Markdown, HTML, TXT
• Text Splitting: Recursive character, token-based
• Vector Stores: In-memory, FAISS, ChromaDB, LanceDB
• Retrieval: Similarity search, MMR diversity

💾 Memory System

• Session Memory: Multi-session management
• Semantic Memory: Vector storage, similarity search
• Self-Learning: Pattern recognition, preference learning
• Closed-Loop: Continuous improvement from interactions

---

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│                     DATA-AI CLI                              │
├─────────────────────────────────────────────────────────────┤
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐         │
│  │   Intent    │  │  Reasoning  │  │    Skill    │         │
│  │ Recognition │  │   Engine    │  │   System    │         │
│  └─────────────┘  └─────────────┘  └─────────────┘         │
├─────────────────────────────────────────────────────────────┤
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐         │
│  │   Code KG   │  │   Memory    │  │   Tool      │         │
│  │   (CKG)     │  │ Closed-Loop │  │Orchestrator │         │
│  └─────────────┘  └─────────────┘  └─────────────┘         │
├─────────────────────────────────────────────────────────────┤
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐         │
│  │  Streaming  │  │   Polish    │  │   Sandbox   │         │
│  │   Output    │  │   Engine    │  │   Security  │         │
│  └─────────────┘  └─────────────┘  └─────────────┘         │
├─────────────────────────────────────────────────────────────┤
│              Multi-Agent Collaboration Layer                 │
│     (Coordinator + Communication + Consensus + Workflow)    │
└─────────────────────────────────────────────────────────────┘

---

📖 Documentation

• Installation Guide
• Configuration Guide
• Model Providers
• Architecture Overview
• API Reference
• Contributing Guide

---

🎯 Usage Examples

Basic Usage

from data_ai import create_client

# Create a client
client = create_client()

# Execute a task
response = client.execute("Analyze the sales data in data.csv")

Advanced Usage with Intent Recognition

from data_ai.core import IntentRecognizer, IntentRouter

# Recognize intent
recognizer = IntentRecognizer()
intent = recognizer.recognize("Search for Python tutorials")

# Route to appropriate handler
router = IntentRouter()
result = router.route(intent)

Multi-Agent Collaboration

from data_ai.core.agent import CollaborationSession, AgentRole

# Create collaboration session
session = CollaborationSession()
session.register_agent("coder", AgentRole.WORKER)
session.register_agent("reviewer", AgentRole.REVIEWER)

# Execute collaborative task
result = session.execute("Implement a sorting algorithm")

---

🛡️ Security

• Sandboxed Execution: Isolated environments for code execution
• Permission System: Fine-grained access control
• Resource Limits: CPU, memory, time constraints
• Policy Management: Strict, moderate, permissive presets

---

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

• Code of Conduct
• Development Guide
• Coding Standards

---

📄 License

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

---

🙏 Acknowledgments

• Inspired by Hermes Agent, Trae, Codex CLI
• Built with LangChain and LangGraph

---

📞 Support

• GitHub Issues
• Discussions

---

⭐ Star us on GitHub if you find this project helpful!