BiasLens

Media bias analysis engine. Shows evidence before conclusions, separates fact from framing, measures uncertainty, and surfaces missing perspectives — every score explainable, reproducible, and challengeable. The "Bloomberg Terminal for media analysis."

Bootstrapped by FleetCrown from the BiasLens Product & Engineering Specification.

Vision

Grow from media into a general-purpose reasoning and narrative-analysis engine — scientific papers, speeches, parliamentary debates, podcasts, video, books, campaigns, referendum info, AI-generated content — the foundation of the Argument Arena vision.

Product philosophy (guides every decision)

• Never tell the user what to think.
• Show evidence before conclusions.
• Measure uncertainty.
• Separate facts from interpretation, and bias from factual accuracy.
• Every score must be explainable; every conclusion reproducible; every article challengeable; competing interpretations coexist.

Stack

Next.js + TypeScript + Prisma/Postgres. Multi-agent pipeline for crawl/extract/score. FleetCrown for orchestration + loops; OrangeCat for funding

• public transparency.

Architecture (target)

11 specialized agents — Crawler, Extraction, Claim, Narrative, Bias, Evidence, Comparison, Visualization, Research, Review, Publishing. Core entities: Outlet, Article, Claim, Evidence, Narrative, Frame, Homepage Snapshot, Editorial DNA. Two loops: hourly Homepage Loop (crawl → extract → analyze → compare → publish) and on-demand Article Loop (extract → analyze → report → feedback).

Roadmap

1. Define the core data model — Outlet, Article, Claim, Evidence, Narrative, Editorial DNA as Prisma schemas (this scaffold seeds prisma/schema.prisma). ✅ DONE (0d46066) — App Router root + initial Prisma migration applied; build, lint, and tsc all green.
2. Build the Extraction + Claim agents to parse an article into structured claims.
3. Bias + Evidence scoring engine — every score explainable.
4. Hourly Homepage Loop crawling + comparing outlets.
5. Counterfactual headline generator + Missing Perspectives.
6. Public dashboard + OrangeCat transparency page.

Getting started

npm install
cp .env.example .env   # set DATABASE_URL
npx prisma migrate dev
npm run dev

Definition of done

Every score documents its input, algorithm, evidence, confidence, and failure modes; every conclusion is reproducible and every article challengeable. tsc + lint + tests pass, and the OrangeCat transparency/funding page is updated on each release.