describe the video, pick the platform, get a layered set built to each platform's own limits, broad category, niche community, clip-specific. no thirty-tag dumps, no #fyp, no invented trend numbers.
classification pass
live previewa quiet morning coffee routine
#coffeetok
#homebarista
#v60recipe
the set appears here, one tag per line, each labelled broad / niche / specific so you can see what it classifies, and swap the ones that do not match your account.
the tools that hand you thirty tags are optimising their own conversion, not your distribution. every platform reads hashtags as a classification input, and a diluted input gets a vaguer audience.
tiktok
3-5
hashtags, in the caption, after the hook line
hashtags are a classification signal, not a push button. three to five tight, specific tags beat a wall of them, extra tags dilute what the video is about.
instagram reels
5-8
hashtags, in the caption or the first comment
reels tolerate a slightly wider set because discovery leans on topic clusters, five to eight, mixing your niche with the specific subject of the clip.
youtube shorts
3-5
hashtags, in the title (1-2) and the description
youtube penalises over-tagging: past 60 tags every hashtag on the video is ignored, and creators report dilution far earlier. stay at three to five topical tags and skip #shorts.
and the tags that do nothing
#fyp, #foryou, #viral, #trending, #explorepage, #shorts, every video wants those things, which is exactly why they carry no information. they describe an intention, not a subject, so this generator will not output them.
the labels on every generated tag map to these three layers, so you can read the set instead of trusting it, and swap any tag that does not match the account posting it.
the category you compete in
one tag that tells the system which shelf this video belongs on, a real category, not #viral. it gives the video a starting audience to be tested against.
the audience you actually want
one or two tags naming the sub-community that would save this. smaller pools convert better because the people inside them already care about the subject.
what happens in this clip
the tags only this video could carry, the product, the place, the technique, the format. these are what make the classification honest and repeatable.
no live trending data, and nobody else has it either
no public api reports today's hashtag volume on tiktok, reels or shorts. tools quoting a 3.8× reach boost are quoting their own landing page. this set comes from your topic plus documented platform behaviour.
hashtags decide the first audience, not the outcome
they classify the video so it gets tested against the right people. whether those people stay is decided by the first three seconds, the cut rhythm and the payoff.
a generated set is a starting point, not a caption
keep the tags that describe your account honestly, drop the ones that do not, and reuse the set across a format instead of reinventing it per post.
if reach is flat, the problem is usually the video
that is the part we actually work on: the analyzer takes a real post apart frame by frame and tells you which beat lost the audience.
three to five on tiktok and youtube shorts, five to eight on instagram reels. this generator caps itself at those numbers on purpose, youtube documents a penalty for over-tagging, and on tiktok extra tags weaken the classification signal instead of widening reach.
because they carry no information. every platform's discovery system already knows a video wants views; tags like #fyp, #viral or #explorepage describe an intention, not a subject, so they do nothing for classification and make the caption read like spam.
no, and we say so rather than implying otherwise. the output comes from your topic plus 2mv studio's platform rules, category, sub-community and clip-specific layers. real trend movement is something we track inside research on the videos themselves, not on tag volume.
discovery works differently on each. tiktok classifies sound-led, trend-adjacent content and rewards tight topical signals. reels leans on topic clusters and saves, so a slightly wider set helps. shorts reads the title and description as search text, which is why one or two tags belong in the title.
tiktok: caption. reels: caption or first comment, both are indexed, so pick whichever keeps the caption readable. shorts: one or two in the title plus the rest in the description, since the title is what search matches against.
no. hashtags decide who gets shown the video first; the first three seconds decide whether they stay. if retention is the problem, run the video through the ai video analyzer and fix the hook before touching the caption.
Start getting full visibility into what makes content go viral.