RISHABH KUMAR commited on
Commit
0c2b47c
Β·
unverified Β·
1 Parent(s): be6f5b1

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +84 -0
README.md ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Intelligent AI Search Engine Agent πŸ€–πŸ”Ž
2
+
3
+ An intelligent agentic search engine powered by **Groq's Llama3-8b**, Streamlit, and multi-source search tools (DuckDuckGo, Arxiv, Wikipedia). Built with LangChain for robust AI agent orchestration.
4
+
5
+ ![Demo](https://via.placeholder.com/800x400?text=Streamlit+Interface+Demo) *(Replace with your demo GIF)*
6
+
7
+ ## Features ✨
8
+ - **Multi-Source Search**: Integrates 3 powerful tools:
9
+ - πŸ¦† DuckDuckGo (web search)
10
+ - πŸ“š Arxiv (scientific papers)
11
+ - 🌐 Wikipedia (encyclopedic knowledge)
12
+ - **Groq-Powered AI**: Utilizes Groq's ultra-fast Llama3-8b model
13
+ - **Smart Agent Optimization**:
14
+ - Auto-stop when sufficient information is found
15
+ - Max 3 iterations to prevent infinite loops
16
+ - Concise answer prioritization
17
+ - **Streamlit UI**: User-friendly interface with chat history
18
+ - **Error Handling**: Robust parsing and API error recovery
19
+
20
+ ## Getting Started πŸš€
21
+
22
+ ### Prerequisites
23
+ - Python 3.8+
24
+ - [Groq API Key](https://console.groq.com/)
25
+ - `.env` file for API keys (optional)
26
+
27
+ ### Installation
28
+ ```bash
29
+ git clone https://github.com/yourusername/ai-search-agent.git
30
+ cd ai-search-agent
31
+ pip install -r requirements.txt
32
+ ```
33
+
34
+ ### Setup
35
+ 1. Get Groq API key from [Groq Console](https://console.groq.com/)
36
+ 2. Create `.env` file:
37
+ ```env
38
+ GROQ_API_KEY=your_key_here
39
+ ```
40
+
41
+ ### Run the App
42
+ ```bash
43
+ streamlit run app2.py # Recommended: optimized version
44
+ # or
45
+ streamlit run app.py
46
+ ```
47
+
48
+ ## Usage πŸ’‘
49
+ 1. Enter Groq API key in sidebar
50
+ 2. Ask questions in natural language:
51
+ - "Explain quantum computing basics"
52
+ - "Latest research on LLM optimization"
53
+ - "Compare CNN and RNN architectures"
54
+ 3. Watch the agent intelligently combine search sources!
55
+
56
+ ## Code Versions πŸ”„
57
+ - **`app.py`**: Base implementation with custom DuckDuckGo search
58
+ - **`app2.py` (Recommended)**: Enhanced version with:
59
+ - Official DuckDuckGoSearchRun tool
60
+ - System instruction prompts
61
+ - Limited iterations (3 max)
62
+ - Early stopping mechanism
63
+ - More robust error handling
64
+
65
+ ## Deployment ☁️
66
+ Project is Deployed at Streamlit Community Cloud:
67
+
68
+ 1. Go to [Streamlit Console](https://share.streamlit.io/)
69
+ 2. Enter your Groq API Key
70
+ 3. Search your query with Intelligent AI Search-Engine
71
+
72
+ ## Contributing 🀝
73
+ PRs welcome! Please follow:
74
+ 1. Fork the repository
75
+ 2. Create feature branch
76
+ 3. Submit PR with detailed description
77
+
78
+ ## License πŸ“„
79
+ MIT License - See [LICENSE](LICENSE)
80
+
81
+ ## Acknowledgments πŸ™
82
+ - Groq for LLM infrastructure
83
+ - LangChain team for agent framework
84
+ - Streamlit for UI components