Releases
FinFram Screener
FinFram Screener is a modular CLI tool for automated stock analysis having a dual architecture: Algorithms + AI forecasting. It pulls historical market data, builds structured prompts, and queries Together.ai’s Mixtral/Groq model to generate forward-looking insights, ideal for due diligence, portfolio screening, or macroeconomic context.
Step 1 - AI Models
The given version is powered by Groq API, so please get your own API key from Groq to get this tool running!
Step 2 - Installation
FF Screener can be downloaded bellow via Github, please follow the instructions and don't forget about the API and .env file
About .env and API implementation
The .env file will be added by you. (new file - set the name as ".env" in the parent directory)
Inside the .env file add
GROQ_API_KEY=INSERT API
INSTERT API is your api key obtained from groq (step 1)
Github/FinFram-Screener
Answers
FF Screener runs by using Groq API's, if you use any other API key, e.g OpenAI, this tool can't run that api natively, you need to adapt the code in order to make it work on other AI models.
This tool is not an AI, it uses a dual architecture, made by Algorithms and the AI model.
Algorithms: FinFram uses standard Python-based algorithms to:
Pull historical stock data (via yfinance)
Calculate metrics like average price, volatility, and latest close
Format structured prompts for forecasting
AI Model: It then sends those prompts to groq's model, a large language model accessed via the compatible API. This model generates forward-looking insights, such as market sentiment or macroeconomic forecasts, based on the structured financial data.
Why This Matters
The algorithmic layer handles deterministic, rule-based tasks (data fetching, math).
The AI layer interprets and extrapolates patterns, offering probabilistic insights that go beyond what traditional screeners can do.
FinFram is a hybrid screener that combines algorithmic analysis with AI-powered forecasting.