Arabic media analysis · Powered by ML and LLMs

Read the news.
Without the bias.

Bayan analyzes Arabic news articles using AI to detect bias, measure neutrality, and surface transparency indicators — instantly.

0
avg analysis time
0
bias indicators tracked
0
point neutrality scale

Why we built this.

Arabic-speaking readers consume news from dozens of sources — regional papers, state broadcasters, independent outlets, and social media. The difference in framing between these sources is vast, but rarely visible to the naked eye.

Bayan was built to make that difference visible. By applying large language models to Arabic text, we can detect the subtle signals of bias that humans often miss: the choice of adjective, the missing quote, the emotional frame wrapped around a fact.

We are a two-person team building a prototype focused on honest, accurate, and explainable analysis. Bayan is not a fact-checker — it is a neutrality mirror.

ProductBayan · بيان
Versionv1.0 — Prototype
Language targetArabic (العربية)
AI engineLarge Language Model
FrontendAstro + React + Tailwind
BackendPocketBase
Team2 members · 2 months

Three steps to clarity.

Bayan processes your article through a pipeline of AI-powered checks, returning a structured report in seconds.

01

Submit an article

Paste any Arabic news article text directly into the analyzer. No account required to get started.

02

AI processes the text

Our LLM analyzes sentiment, loaded language, emotional tone, and framing patterns in the article.

03

Read your report

Get a 0–100 neutrality score, bias indicators, loaded phrases, and an AI-generated Arabic summary.

Everything you need to read clearly.

Built for journalists, researchers, and readers who care about where information comes from.

scoring

Neutrality score

Every article receives a 0–100 score. Higher means more balanced. The score is calculated across multiple dimensions — not just sentiment.

0
moderately neutral · score/100
detection

Loaded language detection

Bayan flags emotionally charged words and persuasive phrases that can subtly shift how a story is perceived — even in neutral-seeming articles.

analysis

Bias indicator mapping

One-sided sourcing, missing perspectives, and selective framing are surfaced individually, so you know exactly where the imbalance comes from.

summary

Arabic AI summary

Every analysis ends with a 1–2 sentence Arabic-language summary of the findings. Designed to be shared, saved, and acted on.

arabic-first

Built for Arabic — from the ground up

Bayan is not a translation of an English tool. The AI model is prompted with Arabic-language context, RTL text handling is native throughout the UI, and the analysis accounts for the specific rhetorical patterns found in Arabic media.

Try it now.

Paste any Arabic article below. The result is generated in real time.

article.txt
/100
neutrality score
Sentiment
Emotional tone
Bias markers
Loaded phrases
AI Summary

Loaded Language
Bias Indicators