₿ CryptoMind — AI Crypto Market Analyst

A production-ready LangGraph ReAct agent that autonomously calls real-time crypto price and news APIs, reasons over live data and serves a multi-turn conversational UI via Streamlit. Powered by Groq (LLaMA 3.3 & LLaMA 3.1) & HuggingFace (Qwen3 & Zephyr) models, FreeCryptoAPI, and NewsAPI.

Ask about any cryptocurrency — the agent fetches live data, reads latest news, and delivers structured investment insights.

This isn't just a chatbot — it's a reasoning agent that:

➡️ decides which tools to use
➡️ gathers real-time information
➡️ compares assets intelligently
➡️ explains results clearly

🏗️ Architecture

User Input
    │
    ▼
Streamlit UI (app.py)
    │
    ▼
LangGraph ReAct Agent (agent.py)
    │  ┌─────────────────────────────────────┐
    ├──▶  crypto_list_tool  → FreeCryptoAPI  │
    ├──▶  crypto_data_tool  → FreeCryptoAPI  │  tools.py
    └──▶  crypto_news_tool  → NewsAPI        │
       └─────────────────────────────────────┘
    │
    ▼
Models (reasoning + response)
    ├──▶ Groq LLaMA 3.3 70B
    ├──▶ Groq LLaMA 3.1 8B
    ├──▶ HuggingFace (Qwen3-8B)
    └──▶ HuggingFace (Zephyr-7B)
    │
    ▼
MemorySaver (multi-turn conversation memory)

⚡ Local Setup

# 1. Clone / enter the project folder
cd crypto_agent

# 2. Create virtual environment
python -m venv venv
venv\Scripts\activate

# 3. Install dependencies
pip install -r requirements.txt

# 4. Add your API keys
# Edit .env and fill in your keys

# 5. Run
streamlit run app.py

API Keys needed (all free tiers available):

Key	Get it at
GROQ_API_KEY	https://console.groq.com
FREECRYPTO_TOKEN	https://freecryptoapi.com
NEWSAPI_KEY	https://newsapi.org
HF_API_KEY	https://huggingface.co/

🗂 Project Structure

crypto_agent/
├── app.py            # Streamlit web UI
├── agent.py          # LangGraph ReAct agent + memory
├── tools.py          # Crypto data + news tools
├── requirements.txt
├── .env
└── README.md