| Primary input format | Video and audio — paste a YouTube/TikTok/Instagram Reel/podcast link, or run the Chrome/Firefox extension live while you watch | Written news — articles and URLs pasted into its News and Bias Checker, plus a Chrome extension that surfaces a site's bias rating as you browse |
| Per-claim verification against live web sources | Every extracted claim gets a 0-100 credibility score, cross-referenced against live sources with the evidence linked per claim | Reliability is assessed from linguistic markers, quote quality, and source diversity rather than by verifying each factual assertion against the open web |
| Political lean scoring (left–right) | Doesn't measure it — no left/right axis at all. This is a real gap, not a design flourish | Its core competency: a -100% to +100% bias score on a 7-point Extreme Left to Extreme Right scale, with a sentence-level bias breakdown |
| Outlet-level ratings and media bias chart | None. Analyzes the individual piece of content in front of you, not the publisher behind it | Thousands of news sources rated (4,900+ by its own count), an interactive 2D bias-vs-reliability media bias chart, plus politician portrayal and policy-leaning ratings |
| Human expert review layer | Fully automated pipeline — fast and cheap, but no humans in the loop | AI output is complemented by reviews from a politically diverse team of trained analysts — a rigor layer BSmeter does not have |
| Deepfake / AI-generated media signals and fallacy flags | Flags deepfake and AI-generated media signals, plus logical fallacies, and syncs relevant viewer comments alongside the analysis | Not applicable — it's a text-and-outlet bias instrument, so synthetic-media detection is outside its scope |