Irram/IRRAM-financial-advice
0
๐ฆ Autonomous Financial Research & Investment Advisor
A multi-agent AI pipeline (Research Agent โ Risk Agent โ Recommendation Agent) built on Groq-hosted LLMs, with a Gradio UI for running on-demand or scheduled equity research.
Setup
- Duplicate/clone this Space.
- In Settings โ Repository secrets, add the keys from the table below.
- The Space will build automatically from
requirements.txtand launchapp.py.
๐ Config Keys Required
Complete reference โ use this to re-add anything if a secret ever gets deleted or you're setting up a fresh copy of this Space. Spelling matters exactly as shown.
Without the two optional keys, the app still works fully โ chart and technical indicators just show "N/A" (with a specific explanation) on tickers Finnhub's free tier doesn't have candle data for, instead of a fallback filling the gap.
Files
app.pyโ main Gradio application (data collection, agents, orchestrator, UI)requirements.txtโ Python dependenciesREADME.mdโ this file (also configures the Space via the YAML front matter above)
Notes
- This app calls the Finnhub REST API for market data (quotes, company profile, basic financials, news) and the Groq OpenAI-compatible endpoint for LLM inference. Scheduled jobs (via APScheduler) only run while the Space is awake โ on free-tier Spaces that sleep after inactivity, scheduled jobs will pause when the Space sleeps and resume on the next wake-up.
- All generated analysis is for informational/educational purposes only and is not financial advice, as stated in the app's own disclaimers.
- Finnhub's free tier restricts historical daily candles (used for the 30-day chart, volatility, and drawdown) to paid plans for most exchanges. The app degrades gracefully when that data isn't available โ it shows "N/A" for those fields and skips the chart rather than failing the whole report. Quotes, company profile, basic fundamentals, and news all work on the free tier.
