Case study · 2026 · AI · Web
You usually need four sources to understand one story.Built an AI-scored news reader. Every article gets a fairness score, so the reader sees right away how slanted it is.
The Truth AI News aggregates news across multiple sources, runs an AI fact-checker on each, and publishes a per-article bias + opinion + intent score so the reader sees how slanted a piece is at a glance. The studio shipped the ingest pipeline, the scoring engine, and the reader-facing surface.
- Next.js
- TypeScript
- Prisma
- PostgreSQL
- OpenRouter
Inside the build
Overview
- n
- Sources cross-checked
- 3
- Score axes per article
- Live
- Cross-source diff
- Postgres
- Persisted history
Pulls news from multiple sources, uses AI to verify facts, summarises, and tells you how objective, how opinionated, and what the piece was trying to achieve. The first portal where editorial bias is visible too.
What shipped
What it does
- Multiple sources on one page
- AI fact-checking
- Bias and opinion scoring
- Summary: what the article was trying to do
The problem
- News is often inaccurate or deliberately slanted
- No time to cross-read four sources
- No way to know how opinionated a piece is
Why it matters
- Fact-checked news in one place
- Bias visible on every article
- Save time · the point is highlighted
How it shipped
- 01 · BRIEF
Bias is a measurement, not an opinion.
Defined the three score axes (bias, opinion-share, intent), trained a scoring rubric on a labelled corpus before the pipeline was built · the LLM scores against a fixed rubric, not vibes.
- 02 · BUILD
Ingest → cross-source diff → score → persist.
Prisma + Postgres schema covers articles, scores, and source-pair diffs. OpenRouter routes to the model best-fit per article-length and language. Scores stored per article version · history visible.
- 03 · SHIP
Reader sees the score before the article.
Article cards lead with the bias / opinion / intent ribbon · the reader knows the slant before clicking through. Cross-source diff visible on tap.
Stack
Ingest
Multi-source crawler
Pulls articles, deduplicates, normalises into a common schema before scoring.
Scoring
Three-axis rubric · bias / opinion / intent
Calibrated against a labelled training corpus · drift watched per release.
Diff
Cross-source comparison
Where two outlets disagree on the same story, the diff is surfaced inline.
Reader UI
Score visible before the article
Article cards lead with a coloured ribbon for each axis · the reader sees the slant first.
Case study
“We wanted to be the news site that doesn't pretend everyone is being honest. They built a system that scores every article, how factual, how opinionated, how political, and shows the comparison side by side. Readers actually click on the score to see why the article got it. People stay on the site twice as long now.”
More work
Next project
Tippsziget