Search Feud

A half-typed search appears — _“why do dogs __” — and behind it sit the six completions people actually search for. The room races to name them. The top completion pays the most, the deep cuts pay the least, and the first chatter to a row takes it whole.

The answers aren’t opinions and they aren’t your chat’s: every board is real search autocomplete, captured and filtered ahead of time. Which means the winning guess is usually the dull one.

Gameplay

The boards

Nearly every board holds six completions, ranked by how often people search them. Points follow the rank, not a survey count:

• The top row is worth 2.5× the bottom one. A lucky late grab never beats reading the room, so it pays to call the obvious completion fast.
• First to name a row claims it. Once taken it’s locked to that player; later guesses of the same completion don’t score.
• Matching is forgiving. Close phrasings and small misspellings land on the right row. See Answer matching.

Type only the ending

The single thing chat gets wrong. The stem is already on screen — send the part that completes it, not the whole search.

Wrong guesses cost nothing and never eliminate anyone.

Configuration

Topic. Scope the boards to one subject, or leave it on Mixed to draw from the whole deck. A topic only appears in the picker when it holds enough boards to fill a session without repeating, so the list is shorter than the full set of subjects in the deck.

Rounds. How many searches make up the game. Scores accumulate across all of them.

Gift claims. See Gifts below.

Layout

1. Marquee — game title and how-to hint.
2. Gift claims — which gift buys which row, when configured.
3. Query bar — the current half-search, with the round’s progress.
4. Suggestions — the completions. Claimed rows fill in with the player who took them and the points banked; the rest stay redacted.
5. Leaderboard — players ranked by total points across the rounds.

The game moves through Lobby → Round → Revealed → Finished.

Gifts

A gift claims a row outright, exactly as if the gifter had typed it first. The defaults are priced so the bigger gift buys the better row:

Any gift can be mapped to any row up to 8, and clearing the picker turns the mechanic off entirely. It’s competitive: a gifted row is one fewer for the room to race for.

Scoring

• Claiming a row banks its full point value; higher rows are worth more.
• Scores accumulate across every round; the highest total wins, and ties are shipped as co-winners.
• Points feed the per-streamer tally behind avatar ring tiers. Wins are tracked per streamer.

Achievements

50 Search Feud wins unlocks the game’s avatar motif.

Where the answers come from

Every board is baked from real search autocomplete ahead of time, then filtered — it is never fetched during a game. That’s deliberate on three counts: the same board plays identically on every stream, nothing unvetted can appear mid-broadcast, and autocomplete’s answers otherwise drift with the news cycle and with wherever the request came from.

The filtering removes suggestions that were true only where or when they were captured — locations, dates, whatever spiked that day — along with not-safe-for-stream content, and drops boards whose rows were six ways of saying one thing.

Search Feud is a parody, and an unaffiliated one. LiveHub is not connected to Google LLC, and uses neither its name nor its logo, wordmark, or colours in the game. “Google” and the Google logo are registered trademarks of Google LLC.