Apppublic

dev-2106/Conversational-AI-Agent

sourceHugging Faceapache-2.0updated 7mo agoView on Hugging Face

0likes

App README

🔎 LangChain - Chat with Search

An AI-powered chatbot that can search the web, query Wikipedia, and find research papers on Arxiv — all in real time using Groq LLM and LangChain Agents. Free to use!

📌 What Does This App Do?

This app is a smart conversational agent that can:

—🌐 Search the web using DuckDuckGo for current information
—📚 Search Wikipedia for factual and encyclopedic knowledge
—🔬 Search Arxiv for academic research papers
—💬 Maintains chat history across the entire conversation
—🧠 Shows its thinking process — watch the agent reason step by step in real time

🛠️ How It Works

User asks a question
        │
        ▼
Groq LLM (llama-3.1-8b-instant) decides which tool to use
        │
        ├──► DuckDuckGo Search  → for general web queries
        ├──► Wikipedia          → for factual knowledge
        └──► Arxiv              → for research papers
        │
        ▼
Agent reasons through results (ReAct framework)
        │
        ▼
Final answer streamed back to user via Streamlit

The agent uses the ZERO_SHOT_REACT_DESCRIPTION strategy — it reads the tool descriptions and decides on its own which tool best answers each question.

⚙️ Tech Stack

Technology	Purpose	Cost
Groq	LLM inference (llama-3.1-8b-instant)	Free
LangChain	Agent and tool orchestration	Free
DuckDuckGo Search	Real-time web search	Free, no API key
Wikipedia API	Encyclopedic knowledge	Free, no API key
Arxiv API	Research paper search	Free, no API key
Streamlit	Web UI with streaming	Free

✅ Only requires a free Groq API key — no other API keys needed!

🚀 How to Use

On HuggingFace Spaces:

1.Open the app
2.Enter your free Groq API key in the sidebar (get one here)
3.Type any question in the chat input
4.Watch the agent search and reason in real time!

Run Locally:

1. Clone the repo:

bash

git clone https://huggingface.co/spaces/dev-2106/langchain-chat-search
cd langchain-chat-search

2. Install dependencies:

bash

pip install streamlit langchain langchain-groq langchain-community duckduckgo-search wikipedia arxiv python-dotenv

3. Create `.env` file:

env

GROQ_API_KEY=your-groq-api-key

4. Run the app:

bash

streamlit run app.py

🔑 API Keys Required

Key	Where to Get	Cost
`GROQ_API_KEY`	console.groq.com	Free

DuckDuckGo, Wikipedia, and Arxiv require no API keys at all.

💡 Features

—✅ Real-time streaming — see the response being generated token by token
—✅ Agent thought process visible — watch the agent decide which tool to use
—✅ 3 search tools — web, Wikipedia, and Arxiv all in one chatbot
—✅ Chat history — full conversation context maintained across messages
—✅ No OpenAI needed — powered entirely by free Groq inference

📋 Requirements

txt

streamlit
langchain
langchain-groq
langchain-community
duckduckgo-search
wikipedia
arxiv
python-dotenv

🐛 Common Issues & Fixes

Error	Cause	Fix
`No module named 'langchain_classic'`	Deprecated package	Use `from langchain.agents import initialize_agent`
`DuckDuckGo RateLimit`	Too many requests	Wait 30 seconds and retry
`GROQ 429 rate limit`	Free tier exceeded	Wait 60 seconds and retry
`Parsing error`	LLM response format issue	Already handled with `handle_parsing_errors=True`

📄 License

This project is open source and available under the MIT License.

<p align="center">Built with ❤️ using LangChain, Groq and Streamlit</p>