Hugging Face's logo

Spaces:
optimization-hashira
/
Multi-Agent-Financial-Assitant
Running

App Files Community
Multi-Agent-Financial-Assitant / README.md
optimization-hashira's picture
optimization-hashira
fix
f38f528
preview code
|
raw
history blame contribute delete
1.47 kB
metadata
title: AI Financial Assistant (Single Container)
emoji: ⚙️
colorFrom: red
colorTo: green
sdk: docker
app_port: 8501
pinned: false

AI Financial Assistant - Morning Market Brief (Single Container)

This application provides a voice-interactive morning market brief. It uses several AI agents for:

  • Speech-to-Text (STT)
  • Natural Language Understanding (NLU - simulated)
  • Financial Data API Fetching
  • Web Scraping for Earnings
  • Document Retrieval (FAISS)
  • Data Analysis
  • Language Generation (LLM)
  • Text-to-Speech (TTS)

All services run within a single Docker container managed by Supervisord.

How to Use

  1. The application will start automatically once the Space is built.
  2. Access the public URL provided by Hugging Face Spaces.
  3. Use the Streamlit interface to record your query or upload an audio file.
  4. Click "Generate Market Brief".

Environment Variables (Secrets)

The following secrets must be set in your Hugging Face Space settings for the application to function correctly:

  • FMP_API_KEY
  • ALPHAVANTAGE_API_KEY
  • GOOGLE_API_KEY
  • GEMINI_MODEL_NAME (Optional): Defaults to gemini-1.5-flash-latest if not set.
  • WHISPER_MODEL_SIZE (Optional): Defaults to small if not set.

The FAISS_INDEX_PATH is configured internally to /app/faiss_index_store. Note that in this single-container setup on HF Spaces, this path might be ephemeral unless specific persistent storage is configured for the Space.