10x at most
Code blocks any order above 10x. The AI cannot change this.
Token fees fund an AI trader on Hyperliquid. Watch its real screen, give it tips, and share the profits.
Paper trading now. No real funds yet.
Trades on HyperliquidLaunches on pump.funBridges with Relay
Each trade of the token pays a fee. A share lands in one public wallet, and the TV counts it live.
When enough builds up, code bridges it to Hyperliquid through Relay. No one signs by hand.
It reads the market all day, then longs or shorts. You watch its real screen on the TV.
When profit passes a milestone above the last high, SOL goes out to holders.
The AI opens each coin and changes the timeframe in a real browser, and you see every move. Code watches the market all day, so it never goes stale.
{
"reason": "stop_near",
"coin": "ETH",
"mark": 2264.4,
"position": "long 4x",
"chat_mood": "23 say long",
"action": "decide"
}The first $50 after the team share goes to trading, so the AI can wake fast. Compute is capped at $30 a day and drops when fees drop.
These sit outside the model. No tip and no bad idea can get past them.
Code blocks any order above 10x. The AI cannot change this.
It can trade any perp with deep volume. Thin coins are blocked.
Each position has a stop on the exchange, so it holds even if the AI sleeps.
The AI gets a key to place orders. Moving money is a separate service with no AI in it.
Chat is grouped into a short mood report. The AI reads it, then decides alone.
A 25% fall from the peak stops new trades for 24 hours.
Yes. A real browser runs the real site, and we send it to the TV as video. The AI picks what it shows. Orders go through the API, not by clicks.
You can give tips in chat with no sign-in. It treats them as mood, not orders, and it makes its own call.
It can lose everything. Only profit above the last high is paid out, so a loss must be won back before the next payout.
We take random hidden snapshots between payouts, then pay SOL by average balance. Buying right before a payout does not help.
10% goes to a public team wallet. Compute takes up to $30 a day. The rest goes to the trading fund.
With no money for compute, the AI sleeps and the TV goes to static. Stops stay on the exchange. New fees wake it.
This is an experiment, not financial advice. An AI trader can lose all of its funds, and the token can go to zero.