Frequently Asked Questions

PressGrade uses AI-powered web research to evaluate media entities across five weighted criteria: Factual Accuracy (highest weight), Correction Culture (high weight), Source Transparency (medium weight), Content Integrity (medium weight), and Editorial Independence (lower weight). Each criterion is scored 1-5, then converted to a weighted percentage score out of 100.

Scores are cached on a monthly basis. The first time a figure is searched in a given month, PressGrade researches them across all five criteria using live web data and generates a fresh score. That score is then cached for the rest of the month, so subsequent searches for the same figure return instantly with a consistent result. At the start of each new month, caches reset and the next search for any figure generates a fresh assessment. If you believe a cached score contains a specific factual error, contact us at hello@pressgrade.io rather than waiting for the monthly reset.

PressGrade's methodology evaluates journalistic practices, not political positions. Our five criteria apply identically to every figure and outlet we score, regardless of where they fall on the political spectrum. In practice this means a conservative outlet with rigorous sourcing and strong correction culture will score higher than a progressive outlet that buries errors and blurs news and opinion. The score reflects the craft, not the viewpoint. We have no financial relationship with any political organization, media company, or ideological group. Our only revenue is Pro subscriptions.

Scores are based on documented behavior over the past 12 months, not reputation, popularity, or how long someone has been in the industry. A figure with a decades-long reputation for credibility may have developed practices recently that do not hold up to scrutiny. A figure with a controversial public image may actually cite sources carefully and correct errors openly. The criterion breakdown on every result page shows you exactly what drove the score. If a result surprises you, start there.

The confidence rating reflects how much verifiable evidence was available for the assessment. Well-known entities with extensive public coverage will have higher confidence scores. Lesser-known figures or niche outlets may have lower confidence, meaning the score is based on less data. Confidence is not the same as trust. The confidence rating tells you how certain we are in the result, not whether the figure is trustworthy.

Scores regenerate automatically each month with the latest available information. You do not need to request a rescore. Simply search for the entity after the monthly cache reset and you will receive a fresh assessment. If you believe a score contains a specific factual error rather than a difference of interpretation, contact us at hello@pressgrade.io and we will look into it.

PressGrade has three access tiers. Guests can see the overall score, trust tier, key finding, and summary for any figure, limited to 5 searches per month. Free account holders see everything guests see, plus the full breakdown across all 5 criteria with scores for each, also limited to 5 searches per month but with search history saved. Pro subscribers see everything in the free tier, plus evidence citations for each criterion, confidence ratings, red flags, strengths, and have unlimited searches. Pro is $8 per month or $5 per month billed annually at $60 per year.

Go to your Account page and click Manage Billing. This opens the Stripe billing portal where you can cancel your subscription immediately. Cancellation takes effect at the end of your current billing period.

We store your email address, search history (entity names, scores, and dates), and Stripe customer and subscription IDs if you are a paying subscriber. We never store credit card details. We do not sell your data, share it with third parties, or use it for advertising purposes. You can view your search history on your Account page and delete your account and all associated data permanently at any time.

PressGrade was built by David Halperin, an independent product builder who grew frustrated with the inability to quickly evaluate whether a media figure or outlet was operating with genuine journalistic integrity. PressGrade is independently funded and has no affiliation with any media organization, political party, government, or ideological group. We do not accept advertising. Our only revenue comes from Pro subscriptions. That independence is not incidental. It is the point.

PressGrade is designed to score public media figures and news outlets: journalists, news anchors, TV hosts, radio hosts, podcasters, political commentators, newspapers, TV networks, magazines, and news websites. We do not score politicians, athletes, musicians, business executives, or private individuals. If you search for someone outside our scope, our system will let you know. If you believe a score has been generated for someone who should not be in our system, contact us at hello@pressgrade.io and we will remove it promptly.

Our scoring uses Claude, Anthropic's AI, with web search capabilities to research each figure across our five criteria using publicly available, documented evidence from the past 12 months. The AI is given a detailed scoring rubric and instructed to cite specific documented behaviors for each criterion. No scoring system is perfect, including ours. That is why we include a confidence rating with every score. We also update scores monthly as new evidence accumulates. If you believe a specific score contains a factual error, contact us at hello@pressgrade.io.

Because PressGrade measures behavior, not reputation. Reputation is shaped by decades of accumulated perception, political affiliation, cultural familiarity, and audience loyalty. Behavior is what someone actually did in the past 12 months. These two things are often different. A figure with a long-standing reputation for credibility may have developed practices over the past year that do not hold up to scrutiny. A figure with a controversial reputation may actually operate with more journalistic rigor than their critics acknowledge. Our scores reflect the evidence, not the narrative.

Your search history is visible only to you when you are logged into your account. We do not sell, share, or expose your search history to any third party. We do not use your search history for advertising purposes. If you delete your account, your search history is permanently deleted along with all other data associated with your account.

Our current coverage is strongest for US-based media figures and outlets. We can score international figures and outlets where sufficient English-language documentation exists. For figures with limited English-language coverage, confidence ratings will be lower, reflecting the reduced evidence base. We plan to expand international coverage as the platform grows.