Our Mission

Tracking accuracy in American cable news.

Verifiable claims, primary sources, the same standard applied to every figure on the platform.

MediaReceipts is a nonpartisan media accountability research platform that systematically tracks the factual accuracy of individual U.S. cable news personalities over time. Every published verdict is sourced, timestamped, and verified by a human editor before it reaches the public.

Keep the
receipts.
See the record.
📋
What we track
Verifiable factual claims made on air by individual cable news personalities: not outlet bias, not journalistic style, not political opinions. Specific claims, checked against primary sources, tracked over time.
🔬
How we work
AI-assisted claim extraction and research, with human editorial review as the quality gate. Every verdict is batch-processed, cross-referenced, and signed off by an editor before publication. We operate on a deliberate week-to-month lag (quality over speed).
📈
Why it matters
No existing tool provides person-level, longitudinal factual accuracy tracking for the anchors and commentators who shape public understanding every day. Aggregate trends over time tell a story that isolated fact-checks never can.
Core Principles
These commitments govern every decision we make: in claim selection, verdict assessment, and public communication.
01
Nonpartisan
Equal scrutiny across the political spectrum. The same methodology, source standards, and verdict taxonomy apply to every figure on the platform, selected for viewership, claim density, and format diversity, not expected verdict outcome.
02
Transparent
Methodology, scoring criteria, and individual claim assessments are fully public and auditable. Radical transparency in our methodology is the primary defense against bias accusations.
03
Evidence-based
Every verdict is supported by primary sources (official government data, court records, congressional records, peer-reviewed research, or direct official quotes) with timestamps and transcripts attached.
04
Longitudinal
The value compounds over time. Aggregate trends across weeks, months, and quarters matter more than isolated incidents. We track trajectories, not moments.
05
Human-verified
AI assists at scale as a research tool, but human editorial judgment governs all final verdicts. No claim reaches the public database without a human editor's review and sign-off.
06
Deliberate pace
Quality over speed. A week-to-month processing lag ensures every published claim has been properly verified, cross-referenced, and reviewed. We are not a real-time fact-checker.
How the Pipeline Works
From clip nomination to published verdict: a six-step process designed for rigor, not speed.
1
Claim Selection
Claims are currently selected by MediaReceipts editorial staff from monitored cable news programming on tracked figures. Each tracked figure's airtime is reviewed against published selection criteria to identify factual claims worth verifying.

A public nomination portal (for community members and outside researchers to submit specific clips and claims for review) is in development and will launch in a future phase. When live, nominations will be queued, prioritized, and processed through the same review pipeline as editorial-sourced claims, with aggregated nomination data published by network so coverage balance is publicly verifiable.
2
Transcript Ingestion
Source transcripts are obtained from broadcast audio of the tracked program. The primary method is machine transcription of the broadcast itself, producing a speaker-aligned, timestamped transcript. Where machine transcription is unavailable, the Internet Archive TV News Archive is the configured secondary source. Each transcript records how it was obtained, and that provenance will be reflected on a claim's record when an archive-sourced transcript is used.
3
Claim Extraction
AI-assisted processing identifies verifiable factual claims within the transcript, distinguishing them from opinions, editorial framing, and non-falsifiable assertions. Borderline cases are flagged for direct editor review.
4
Research & Verification
Each claim passes through two sequential review stages. Gate 1 assesses whether the claim is a verifiable factual assertion worth checking. Claims that pass receive a full independent research pass at Gate 2, where sources are evaluated and a verdict is assigned. A human editor reviews and signs off at both stages. AI findings inform but do not determine the final verdict.
5
Correction Research
Before publication, every claim's research pass includes a structured correction search: checking the network's correction page, the figure's on-air and social statements, and external fact-check feeds. Any correction found at research time is recorded with the original verdict.

Continuous post-publication correction monitoring, automated re-checking of published claims as new corrections appear in the wild, is in active development. When live, it will surface late-arriving corrections directly on the affected verdict and on the figure's record. Whether a corrected claim was subsequently repeated will become a structured field in the published record.
6
Batch Publication
Verified verdicts are published in batches (weekly or monthly) with full sourcing, transcript excerpts, and decision memos attached. Figure profiles and accuracy trends update automatically on publication.
The Six Inaugural Figures
Selected to demonstrate nonpartisanship from Day 1: the highest-viewership personalities at each major cable news network, regardless of political lean.

We chose figures with the highest viewership at each network, across the political spectrum. The same methodology applies to every figure on the platform. The data will tell the story.

Fox News
Jesse Watters
Jesse Watters Primetime · Mon–Fri 8PM ET
High viewership; frequent specific factual claims on crime, immigration, and economics.
Fox News
Bret Baier
Special Report with Bret Baier · Mon–Fri 6PM ET
Fox's chief political anchor and host of the network's flagship straight-news hour. Solo-host format produces clean diarization and stable claim attribution.
MSNBC
Rachel Maddow
The Rachel Maddow Show · Mondays 9PM ET
Consistently the highest-ranked non-Fox program in cable news' weekly top 15. Known for layered, sourced arguments built across long-form segments, a format that tests the methodology's source-evaluation capabilities.
MSNBC
Jen Psaki
The Briefing with Jen Psaki · Tue–Fri 9PM ET
Former White House Press Secretary; frequent policy-grounded factual claims with traceable public records.
CNN
Anderson Cooper
Anderson Cooper 360° · Mon–Fri 8PM ET
CNN's most-watched program. Long-tenured straight-news anchor with broad coverage producing a high volume of factual claims across policy, breaking news, and international affairs.
CNN
Jake Tapper
The Lead with Jake Tapper · Mon–Fri 5–7PM ET
CNN's chief Washington correspondent hosting a two-hour afternoon block. High volume of policy and political coverage across the daily news cycle.
Our Transparency Commitments
Accountability only means something if the accountants are themselves accountable.
📖
Full public methodology
Verdict definitions, the Gate 1 selection rubric, source standards, the two-gate review process, the appeals process, the intent policy, and the rules for how accuracy figures are calculated are published in full on the Methodology page. Nothing about how we reach verdicts is proprietary or hidden.
✏️
Public appeals process
Any individual (including the figure in question, their representatives, or the public) may submit an appeal. Appeals receive an initial response within 7 days and a substantive editorial review within 30 days. All verdict revisions are logged in a public changelog.
🔁
Correction tracking
Every published verdict is researched for corrections at the time of publication, and any correction found is recorded with the verdict and never erased. As post-publication monitoring infrastructure comes online, late-arriving corrections will be surfaced on the affected verdict and reflected in the figure's record. Whether a corrected claim was later repeated is and will remain a publicly visible field.
⚖️
Nomination balance reporting
Because community nominations can skew in volume, we publish aggregated nomination data by network so external observers can verify that editorial coverage remains balanced across the political spectrum.
Who We Are

MediaReceipts is operated independently by a single researcher with a background in applied AI systems and analytical research. The platform is built and run without institutional affiliation, political sponsorship, or advertising revenue of any kind.

The editorial process follows a published, multi-stage methodology. Every claim passes through structured selection, transcript verification, source review, and verdict assignment under documented rules. As contributor and second-reviewer positions are added, the staffing model and any review-coverage metrics will be reflected on the Methodology page.

MediaReceipts has no advertisers. No tracked figure or their employer has any relationship with, or financial stake in, the platform. Our only obligation is to accuracy.

Independent Advisory Board
Coming in Phase 3
MediaReceipts is recruiting 2–4 independent board members from across the political spectrum (working journalists, academics, and media critics) with published bios and conflict-of-interest disclosures. The advisory board will provide editorial oversight and authorize the release of the Misleading verdict category.
Seating timeline to be announced

Submit an appeal

Dispute a verdict? Any individual or representative may submit a formal appeal for editorial review. All appeals are acknowledged within 7 days.

Submit an appeal →

Get in touch

Press inquiries, research partnerships, data licensing, or general questions: we read every message.

Contact us →