v0.1.0

First public beta release of FairSense-AgentiX, an agentic fairness and AI-risk analysis platform.

What's in this release

FairSense-AgentiX uses agentic AI workflows to analyze text, images, and datasets for bias and risk. It combines a ReAct-style reasoning loop with tools for OCR, vision-language models, embeddings, and LLMs.

Highlights

• Agentic reasoning – ReAct loop with dynamic tool selection and self-critique
• Multi-modal analysis – Text bias, image bias, and AI risk assessment
• Tool ecosystem – OCR (Tesseract, PaddleOCR), VLMs, sentence embeddings, FAISS, and LLM adapters (OpenAI, Anthropic)
• APIs – FastAPI REST API and WebSocket streaming
• Configurable – Swap LLM providers, tools, and models via configuration

Requirements

• Python 3.12+
• 4GB+ RAM recommended for ML models
• OpenAI or Anthropic API key for LLM features

Install

pip install fairsense-agentix