Methodology

The data source

All YouTube data on this site comes from the official YouTube Data API v3. Nothing is scraped from search results, and nothing is bought from a third-party keyword vendor.

The Trending Tags Index is built from one endpoint: videos.list called with chart=mostPopular and a regionCode. That is the same list YouTube shows on its own Trending tab for that country. For each country we page through it until we have up to 200 videos, and we store, per video: the title, the channel, the view / like / comment counts at fetch time, the publish date, the category, the duration, and the video's own snippet.tags array.

Hashtags are not a separate API field. We read them out of the video's title and description with a simple #word match, lowercase them, and count each one once per video.

Countries covered

• India — regionCode=IN
• United States — regionCode=US
• Pakistan — regionCode=PK
• Indonesia — regionCode=ID
• Bangladesh — regionCode=BD
• Brazil — regionCode=BR
• Egypt — regionCode=EG
• Nigeria — regionCode=NG

Refresh schedule

The fetch runs once a day at 03:00 UTC. Every page that shows index data prints the exact timestamp of the fetch it is reading from, so you can always see how old the numbers are. The most recent fetch finished 4 September 2026, 15:04 UTC.

Raw trending videos are kept for 90 days and then dropped. The daily aggregate rows (tag counts, hashtag counts, title statistics) are kept indefinitely.

How each number is worked out

Trending videos (per tag or hashtag)

The number of distinct videos carrying that tag inside the rolling 7-day window. A video that trends on four days in a row is counted once, not four times. Tag comparison is case-insensitive; the label shown is the spelling used by the first video we saw it on.

Avg views

Total views of those distinct videos divided by their count. Views are recorded at fetch time, so this is "how many views these videos had while they were trending", not a lifetime figure and not a prediction.

Title patterns

Measured across every trending title in the window. "Contains a digit" is a regex for any numeral. "Phrase in brackets" matches round, square or curly brackets with 1–40 characters inside. "Question form" is a literal question mark. "All-caps word" is any run of three or more capital letters not followed by a lowercase letter. The typical-length band is the 25th to 75th percentile of title length in characters. The 25 top words exclude a fixed list of stop words.

Title Score (0–100)

The score on the title generator is a similarity benchmark, not a quality judgement and not a ranking prediction. Six traits are checked against your title. Each trait is weighted by how common it is among the titles currently trending in the country you picked — so in a country where 78% of trending titles contain a digit, having a digit is worth more than it is in a country where 12% do. The score is the share of the available weight your title earns:

score = 100 × Σ(prevalence of trait × trait present) ÷ Σ(prevalence of trait)

Nothing in that formula is a guess. Every prevalence figure is measured from the live index and shown next to the factor it belongs to. A score of 100 means "your title looks exactly like what is trending in that country today", which is useful information and is not the same as "this title will do well".

Evidence labels on generated tags

When a generator suggests a tag, we look it up in the index for the country you selected. An exact match reports the real video count and average views. A tag that only appears as part of a longer index tag is reported as a partial match. Anything with no match at all is labelled untested — we are telling you the index has never seen it, not that it is bad.

Quota limits

The YouTube Data API charges a project 10,000 quota units per day by default. A videos.list call costs 1 unit no matter how many results it returns, so a full index rebuild across 8 countries costs about 32 units a day. Interactive tool lookups cost 1 unit each and are cached for 24 hours per video or channel id, so repeat lookups of the same URL are free.

If the quota is ever exhausted, the tools say so plainly and stop — they do not fall back to invented data.

What these tools cannot do

• They cannot show tags for a private, deleted or age-restricted video. The API returns nothing for those and so do we.
• They cannot show tags the uploader did not set. Many large channels leave tags empty; an empty result usually means an empty field, not a failure.
• They cannot tell you search volume. YouTube does not publish it. Nobody who claims to have it is reading it from YouTube.
• They cannot predict whether a video will rank or trend. The index describes what is already trending; it is a description, not a forecast.
• The index is not a census of YouTube. It covers the Trending chart of 8 countries — a few hundred videos per country per day, weighted heavily toward music, film and gaming, because that is what trends.
• Earnings figures are estimates. The money calculator multiplies public view counts by a CPM range you choose. Real payouts depend on the advertiser mix, watch time, country and format, none of which are public.

Corrections

If a number here looks wrong, tell us and we will check it against the raw snapshot. Fixes are written up with a date on the changelog. Contact details are on the about page.

Every one of these reads live data from the YouTube Data API or from the Trending Tags Index. No account needed.