⬡ Artha Backend

REST API for the Artha AI financial analyst. Built with FastAPI + LangGraph + Groq.

Artha is an AI Agent built for Financial Assitance with various capabilities

• Data Fetching
• Data Analyses
  • Technical Analyses
  • Fundamental Analyses
• Visualization
• Predictions,
• Education,
• Summarization

Two Approaches:

• Single Agent
• Multi-Agent
  • A guide/router that decides which sub agent to use.
  • An analyst tool to perform stock data related tasks.
  • Aggregator tool to fetch and summarize data from internet, news, and files.

---

Sample Output

Stack

Aspect	Tool/Source
API	FastAPI + Uvicorn
Agent	LangGraph ReAct + LangChain
LLM	Groq — Llama 3.3 70B
Stock Data	yfinance (NSE + BSE)
Web Search	Tavily
News	NewsAPI
Forecasting	Amazon Chronos T5 Tiny
Document RAG	ChromaDB + SentenceTransformers

---

Setup

git clone https://github.com/Ishan2608/Artha.git
cd artha-backend
python -m venv venv
source venv/bin/activate        # Windows: venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env            # fill in your API keys
uvicorn main:app --reload

API runs at http://localhost:8000 · Swagger at http://localhost:8000/docs

---

Environment Variables

GROQ_API_KEY=
GEMINI_API_KEY=
TAVILY_API_KEY=
NEWS_API_KEY=
UPLOAD_DIR=uploads

---

API

Method	Route	Description
POST	/chat	Send a message, get agent reply + optional chart data
POST	/upload?session_id=	Upload a file (PDF, DOCX, XLSX, CSV, TXT, PPT)
POST	/context	Inject raw text context into a session
DELETE	/session/{id}	Clear session and delete uploaded files
GET	/session/{id}/files	List files in a session
GET	/health	Health check

---

Testing

python tests/scripts/test_tools.py   # tool-level tests, no agent
python tests/scripts/test_agent.py   # automated end-to-end, 8 prompts
python tests/scripts/test_run.py     # interactive terminal chat

Logs saved to tests/logs/.

---

Structure

├── main.py              # FastAPI routes
├── agent.py             # LangGraph agent + all tool definitions
├── config.py            # Env config via pydantic-settings
├── tools/               # Stock data, web/news search, ticker lookup, forecasting
├── utils/               # Doc parser, RAG engine, session store, formatters
├── models/schemas.py    # Pydantic request/response models
├── data/listings/       # INDIA_LIST.csv — merged NSE + BSE listings
└── tests/               # Tool tests, agent tests, terminal chat client

Disclaimer

For educational purposes only. Nothing produced by this API constitutes financial advice.