A new prediction game for the questions you're already arguing about. Make a Call. The scoreboard remembers. The internet finally has a record of who's actually right.
Live receipts on @calldhq · Week 01 call is up
The Call · Week 01 · Live now
Tesla quietly removed the “1H 2026” timeline from those five cities in the Q1 deck. $25B capex is going to factories, AI, and Optimus — not Robotaxi expansion. Musk on the call: Robotaxi revenue won’t be material in 2026.
See the call on X →New Call drops every Monday. Right or wrong, the receipt gets posted.
How it works
Anyone can post one. "Will Trader Joe's open in Williamsburg by August?" "Will The Bear win an Emmy?" "Will my friend's startup still exist in a year?" Each one has a clear answer eventually.
Yes or no. The more confident you are, the more points you risk. Odds shift in real time as the crowd weighs in — like a market for opinions, except the prize is being right.
When the date arrives, our AI checks every public source it can find and resolves the question. Right calls climb. Wrong ones get logged. The scoreboard updates. Forever.
What's coming
Once we open up, anyone will be able to make Calls, build a record, and climb the scoreboard. Until then, we post one weekly Call publicly — same rules, same receipts.
Public weekly Calls. Posted on @calldhq every Monday. Verified on the date specified. No edits, no quiet deletes.
Open beta for the first 1,000. Founder-badge holders get in before everyone else. The badge is permanent.
The full game. Make your own Calls, vote on others, climb the scoreboard, build a record the internet can't forget.
Loud, confident, wrong people sound the exact same as quiet, careful, right people. Scoreboard doesn't lie.
— Calld, founding principle
Why now
What used to take a human moderator now takes an agent thirty seconds and a tenth of a cent. That single change makes a million tiny questions economically possible to track.
Polymarket did billions in volume last election. Kalshi made event contracts legal. The category is wide open — and almost nobody is building for the questions that actually matter to your life.
The internet has scoreboards for everything except the one that should matter most: who actually calls things correctly. We're building it. Receipts, forever.
Questions you'll ask
No. Calld is points-only — no real money in, no real money out. Legally it's a game, the same way fantasy football and Manifold Markets are games. We have no plans to change that without a serious regulatory plan in place.
When a question's resolution date arrives, an AI agent checks public sources — news, official websites, Google Maps, Wikipedia, social media. For ambiguous outcomes, a panel of top-ranked human predictors gets the final call. We publish every resolution with sources.
None. We're free during beta because we want the best predictors in the world on the platform first. Eventually we may charge for premium features (better analytics, custom leagues, etc.) — but the core game stays free.
Sports betting is regulated separately and aggressively in the U.S. We're skipping it for V1 to keep our legal story clean. Plenty of room without it.
A small team building Calld in public. The most reliable way to see what we're up to in real time is @calldhq on X — weekly Calls drop there first, with full receipts.
Because every Call is posted publicly before it resolves — no editing, no quiet deletions, no rewriting history. Right or wrong, the receipt goes up. Watch @calldhq for a few weeks and you'll see how it works.
Get on the early access list. First 1,000 handles get permanent founder badges.