Brackify Radar searches X, YouTube, Bluesky, and Brackify Social in one query. It understands 30 fighting games and 900+ characters, so "Juri BnB" finds Juri combos and "Jin punish" finds Jin punish combos, not noise.
Search the way you would ask a friend: "Juri BnB", "SF6 Ken corner combo", "how do I deal with Jin", or just "Guilty Gear Strive". Autocomplete suggests games and characters as you type, and typos like "zangif" still resolve. Every source is on by default. Toggle X, YouTube, Bluesky, or Brackify off if you only want one kind of result, and sort by Relevant, Newest, or Popular. Every result is a combo, so there are no content filters to set. Radar returns everything it already knows within a moment, then refreshes each source independently in the background. New arrivals merge into the list without a reload, and a slow or failing source never blocks the others. One tap on the bookmark saves a result to All Saved and opens a picker where you can file it into up to ten collections such as "Ken oki" or "Tekken 8 punishes". Collections can mix Radar finds, saved Brackify posts, and bookmarked events.
Brackify Radar is a fighting game search engine built into Brackify. It takes one query, works out which game and character you mean, and searches X, YouTube, Bluesky, and Brackify Social for combos at the same time. Results come back as one ranked list of combos you can sort and save. It is free and public at brackify.gg/radar.
Yes. Searching every source is unlimited and free on every plan, and you do not need an account to use it. X posts are collected on a schedule, so there is no per-account X allowance.
Radar ships with a canonical roster of 30 games and more than 900 characters, including Street Fighter 6, Tekken 8, Guilty Gear Strive, Mortal Kombat 1, 2XKO, Fatal Fury: City of the Wolves, Granblue Fantasy Versus: Rising, The King of Fighters XV, Dragon Ball FighterZ, Super Smash Bros. Ultimate, and legacy titles such as Street Fighter V, Tekken 7, and Ultimate Marvel vs. Capcom 3. Aliases like SF6, T8, GGST, MK1, and character nicknames such as Gief, Chun, or Bison are recognized. The roster is data, so new DLC characters are added without an app update.
Three ways. It searches four platforms at once instead of one tab per site. It understands fighting game vocabulary, so a query like "Juri BnB" is read as Street Fighter 6, Juri, combo intent rather than a string of random words. And it keeps an index, so results that were fetched before are available instantly and older X posts remain searchable after the seven-day window of X recent search has passed.
Engagement is normalized per platform before it is compared. A YouTube view count is not weighed against a Bluesky like count on raw numbers; each source is scaled against a baseline typical for that platform. Relevant sort combines entity and text match with recency and normalized engagement, while Newest sorts purely by publication time.
Mostly no. A deterministic parser handles case, punctuation, plurals, aliases, and small typos, and it resolves games and characters against the canonical roster. Only when a query is conversational and the parser is not confident does Radar ask a language model to extract the game, character, and intent, and even then the answer is validated against the roster, so the model cannot invent a character. The rest of the pipeline, including ranking and content classification, is rule based.
Radar collects X posts on a schedule by the hashtags players use for each game and character, the way a search engine builds its index. Each post arrives already labeled with the right game and character, results load instantly, and everyone sees the same X posts. every search still checks YouTube, Bluesky, and Brackify Social live for new combos.
Yes. Radar is a responsive web page and is built into the Brackify iOS and Android apps, where it has its own tab in the bottom navigation. Saved results and collections sync across web and mobile because they live on your Brackify account.