500,000+ Discover articles
Our corpus continuously indexes articles that surface on Discover (FR + EN + IT + DE + PL + NL). We have title, date, publisher, traffic estimate, category. Not a guess — real signal.
Type a category — we scan 500,000+ indexed Discover articles, cross-reference with Google Trends, and surface the top 3 priority topics to publish now with examples of titles that win.
No magic. We combine 4 data sources + a topical clustering model.
Our corpus continuously indexes articles that surface on Discover (FR + EN + IT + DE + PL + NL). We have title, date, publisher, traffic estimate, category. Not a guess — real signal.
We compare the editorial acceleration (articles published on the topic) against the demand signal (Trends volume), filtering out media-only echo chambers and keeping topics where readers actually engage.
Weighted blend: Discover-surface velocity (50%) + search growth (30%) + P90 traffic on the topic (20%). Above 80 = publish now.
Low = fewer than 5 established publishers have covered the angle. Medium = 5-15 publishers. High = ≥15 publishers, you're late. Sweet-spot: low + momentum ≥ 80.
For every topic, we surface the 2-3 best-performing titles in Discover for that thematic cluster. You see exactly which angles work — the promise, the emotion verb, the structure.
For each topic, we surface the optimal day-of-week + hour observed for the Discover peak across similar articles. Publishing 12 hours early ≠ publishing 3 days early.
That's our active corpus — articles indexed in the last 90 days, fresh enough to reflect current Discover behavior. We purge older items that no longer represent the current algorithm. Recency matters more than scale here: 500K fresh articles > 5M articles a year old.
We use an internal prediction model blending public signals (social engagement, persistence in Discover over N days, publisher authority) calibrated against Search Console data shared by our consulting clients. Median and P90 are estimates, not official GSC numbers — but the relative ranking is reliable.
The engine works but we're still calibrating momentum scoring per industry. Same caps as the other free tools: 3 scans / 24h as guest, 10 scans / 24h with a free account. For continuous or batch use, request a Pro access.
All of them — it's free-form topical clustering, not a closed taxonomy. Type pediatric health or European sharing economy or indie games, the engine groups matching articles and runs its analysis. More specific input → better output.
To go beyond the 10 free scans / 24h, a monthly Pro plan unlocks continuous tracking of 5-10 topics + email/Slack alerts when momentum crosses your threshold + API access. Request a quote via /pro-contact.
Start free, scale up when you need to.
Continuous tracking of 5-10 topics, email/Slack alerts, API access, custom integrations: we'll shape the offer with you.
Topic Radar surfaces the angle. Title Lab scores your H1 against 500,000 Discover articles. Image Validator audits your hero. The full trilogy — no detail left to chance.