README.md

---
title: PragmaLens AI
emoji: 🔍
colorFrom: indigo
colorTo: blue
sdk: docker
pinned: true
license: mit
short_description: Linguistic audit for pragmatics and logical fallacies.
---

# 🔍 PragmaLens AI: Intelligent Discourse Audit

**PragmaLens AI** is a professional-grade linguistic analysis engine engineered to deconstruct the mechanics of human communication. By merging **applied linguistics** with state-of-the-art **Large Language Models (LLaMA 3.3 via Groq)**, PragmaLens provides a rigorous audit of textual coherence, intent, and logical validity.

---

## 🎯 The Analytical Framework

PragmaLens transforms raw text into structured, actionable insights using a multi-layered linguistic approach:

### 1. Pragmatic & Gricean Analysis
Goes beyond surface-level text to uncover **conversational implicatures**. The system evaluates adherence to Grice’s Maxims (Quantity, Quality, Relation, Manner), identifying where communication is intentionally indirect or obscured.

### 2. Semantic & Ambiguity Detection
Identifies **lexical and structural ambiguities** that impede clarity, providing a precise audit of how information is framed to avoid—or cause—misinterpretation.

### 3. Logical & Epistemic Audit
Acts as a stress-test for argumentation. It isolates rhetorical fallacies—such as *False Dilemmas*, *Appeals to Authority*, or *Ad Hominem* attacks—highlighting how texts attempt to bypass critical reasoning.

---

## 🌍 Multilingual Philological Precision

Language is the primary lens through which reality is shaped. PragmaLens is optimized for high-fidelity processing across diverse linguistic families, ensuring cultural and pragmatic nuances are preserved.

* **Supported Languages:** English, Arabic, Spanish, Berber (Tamazight), Catalan, French, and Italian.

---

## 🛠️ Technical Architecture

* **Backend:** FastAPI (Python) with high-performance document extraction via `pdfplumber`.
* **Intelligence:** Powered by LLaMA 3.3 70B via Groq for low-latency, high-reasoning output.
* **Frontend:** A clean, minimal UI built with Tailwind CSS, designed for professional workflows.
* **Deployment:** Fully containerized on Hugging Face Spaces.

---

## 🚀 Quick Start

To run PragmaLens locally:

```bash
git clone [https://huggingface.co/spaces/jamalinu/PragmaLens-AI](https://huggingface.co/spaces/jamalinu/PragmaLens-AI)
cd PragmaLens-AI
docker build -t pragmalens-ai .
docker run -p 7860:7860 pragmalens-ai

⚖️ License & Research
© 2026 Jamal Saghraoui. Licensed under the MIT License. This project is dedicated to the advancement of linguistic research and computational pragmatics.