What is Honest Headlines?

Honest Headlines is a news comparison tool that aggregates partisan media coverage, separates facts from opinions, and scores relative accuracy so readers can compare framing without reading only one side.

How do you score accuracy?

Scores estimate how checkable and specific claims appear in available headlines and summaries—numbers, attribution, dates, and quotes raise accuracy; hedges and loaded language lower it. Scores are not definitive truth verdicts.

How is composite trust calculated?

For each source or author we average four 0–100 metrics across stored articles, then compute 35% accuracy + 30% fact ratio + 20% × (100 − sensationalism) + 15% × (100 − opinion ratio), rounded to 0–100. Higher trust means more checkable, less loaded framing in available text—not a permanent verdict or guarantee of truth.

What is the Trust page?

Trust ranks sources or authors from durable article history using fact, opinion, accuracy, and sensationalism averages. Sort by Trust (source) or Trust (author) for a composite score, or by Accuracy, Facts, Opinion, or Sensationalism to rank by that metric alone.

Is the app biased toward one party?

The product is designed for side-by-side comparison. Source lean labels reflect commonly reported audience tilt, not formal party affiliation.

Where do articles come from?

Public RSS feeds from outlets across the political spectrum. Only recent items are shown so the feed stays current.

Methodology · media literacy · common questions

FAQs: how Honest Headlines scores news framing

Answers about fact vs opinion scoring, accuracy, sensationalism, composite trust, quote comparison, bias memory, sources, and privacy.

Honest Headlines is a nonpartisan news comparison tool. This page explains what the product does, how scores are produced from headlines and summaries, what they do and do not mean, and where articles come from. Use it when you need plain-language methodology before trusting a meter on the dashboard or bias tracker.

Short answers cover accuracy heuristics, lean labels, quote comparison, optional AI analysis, overlapping topics, the twelve-month bias memory, and how composite trust is calculated on the Trust page. For hands-on tools, open the live dashboard, overlapping topics, or bias tracker.

Questions and answers

What is Honest Headlines?: Honest Headlines is a news comparison tool that aggregates partisan media coverage, separates facts from opinions, and scores relative accuracy so readers can compare framing without reading only one side.
How do you score accuracy?: Scores estimate how checkable and specific claims appear in available headlines and summaries—numbers, attribution, dates, and quotes raise accuracy; hedges, loaded language, and pure evaluation lower it. Scores are not courtroom truth verdicts.
Is the app biased toward one party?: The product is designed for side-by-side comparison. Source lean labels reflect commonly reported audience tilt, not formal party affiliation. Analysis focuses on language signals rather than endorsing a side.
Where do articles come from?: Public RSS feeds from outlets across the spectrum. Only items from roughly the last twenty-four hours appear in the live feed so the dashboard stays focused on the current cycle.
What is quote comparison?: Quote comparison extracts direct quotations and attributed paraphrases about the same event from both sides, then highlights selective editing, context removal, loaded paraphrasing, omission, or emphasis shifts in the available text.
Do I need an API key?: No. Heuristic scoring works without a key. Adding a SpaceXAI / xAI key enables deeper model analysis for article and topic comparison when you want richer claim breakdowns.
What is the bias tracker?: The bias tracker summarizes fact, opinion, accuracy, and sensationalism signals per source and named author for the current fetch, with an optional rolling twelve-month memory so one headline does not define an outlet forever.
What counts as an overlapping topic?: Topics appear when stories from both sides share at least 25% keyword overlap in titles and summaries. You can open coverage previews and run fact or quote comparisons from the Overlapping topics page.
What is the Trust page?: Trust ranks sources (or authors) from durable article history using fact, opinion, accuracy, and sensationalism averages. Sort by Trust (source) or Trust (author) for a composite score, or by Accuracy, Facts, Opinion, or Sensationalism to rank by that metric alone. Lean filters limit which stored articles are included.
How is composite trust calculated?: For each source or author we average four 0–100 metrics across their stored articles, then compute: 35% accuracy + 30% fact ratio + 20% × (100 − sensationalism) + 15% × (100 − opinion ratio), rounded to 0–100. Higher trust means more checkable, less loaded framing in available text—not a permanent verdict on an outlet or a guarantee of truth.

How Honest Headlines works

1. Aggregate both sides

Public RSS feeds from right-leaning and left-leaning outlets are fetched in parallel. Only recent items (within the last twenty-four hours) enter the live feed.

2. Cluster shared topics

Titles, summaries, named entities, and related wording are compared so stories about the same event can sit next to each other.

3. Score facts vs opinions

Heuristics look for numbers, dates, attribution, and quotes (fact-leaning) as well as modals, tribal labels, and hype (opinion and sensationalism). Accuracy reflects checkability of available text.

4. Compare quotes

When both sides surface remarks about the same people or events, the quote tool lines them up and flags selective shortening, missing context, or emphasis shifts.

What the scores mean (and what they do not)

Fact ratio estimates how much of the scanned language looks like reportable, checkable content. Opinion ratio rises when the text leans on prescriptions, absolutist claims, or emotionally loaded verbs.

Accuracy score is higher when claims are specific and attributed, and lower when hedges or mixed framing dominate. Sensationalism rises with clickbait patterns and heat language—not automatically “false.”

Metrics operate on headlines and RSS snippets first, with optional deeper model analysis when an API key is configured. They are tools for comparison, not courtroom findings.

Sources, lean labels, and privacy

Articles come from public RSS feeds across a wide set of outlets. Lean labels (right-leaning / left-leaning) reflect commonly reported audience or ideological tilt—not formal party endorsements.

Analysis runs on titles and summaries exposed by those feeds. When optional AI analysis is enabled, requests go through a server-side key. Always open the original article before drawing firm conclusions.

Ready to inspect today’s cycle? Open the live dashboard, overlapping topics, or the bias tracker.