HyperAgents: Different Agents. One Arena.

Every market creates a conversation. One trader sees momentum. Another sees an overreaction. A third waits for a clearer trend. They can look at the same chart and reach completely different conclusions. HyperAgents gives those approaches an identity, a shared starting line, and a record that anyone visiting the arena can inspect.

The idea is simple: make strategy something people can explore. An agent should have more than a name and a symbol. It should have a recognizable approach, explicit conditions for acting, and a way to compare its decisions with the underlying market. HYPE is the common reference point. Vector, Rebound, and Anchor are the first three contenders.

The current arena is an interactive preview. It reads public Hyperliquid prices and uses historical HYPE perpetual candles to run deterministic strategy simulations. Hyperliquid documents the public API that supplies those market observations. The live element is the market data; the strategy results are historical simulations. This distinction defines what visitors can actually use today. [Hyperliquid API documentation](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint)

Vector follows momentum. Its rule compares an eight-hour simple moving average with a twenty-hour average. When the shorter average rises above the longer one, the simulated strategy takes a long position. When that relationship reverses or becomes equal, it returns to cash. Vector turns a directional idea into a rule that can be examined candle by candle.

Rebound approaches the same market from a different angle. It looks for a close more than one standard deviation below the rolling fourteen-hour mean, then simulates an entry. It exits when the close returns to or exceeds that mean. Its identity comes from seeking a return toward the recent average. That behavior can work in some sequences and struggle in others; the arena shows the resulting path.

Anchor uses a twenty-four-hour simple moving average as its threshold. A close above that average produces a long signal. A close at or below it produces a cash signal. This is a slower filter than Vector's crossover. The name describes the strategy's character, while its record shows the actual consequences of its rule. No preset is presented as the permanent winner.

All three standard contenders receive the same simulated starting budget, equivalent to 100 HYPE at the first candle's close. They are evaluated over the same seven-day window of closed hourly candles. Each simulation moves between a long HYPE position and cash at one-times exposure. The indicators require an initial warm-up period before they can issue a usable signal.

The sequence matters. A rule checks a candle after it closes, and a resulting simulated order fills at the next candle's open. A modeled fee of 0.05% is applied to each execution. The engine marks equity at subsequent closing prices, and any position still open at the end stays marked to the final close. It does not force an artificial last sale to tidy the result.

Those choices make the comparison understandable. Visitors can inspect USD return, performance relative to holding HYPE, maximum drawdown, modeled fees, and the trade record. USD return and return against HYPE answer different questions. A strategy can gain in dollars while finishing behind a simple HYPE holding position. The arena makes that difference visible instead of compressing everything into one attractive number.

The simulation has defined limits. It does not include spread, slippage, or perpetual funding, and a seven-day historical result does not establish future performance. The purpose of the record is to reveal what happened under the displayed assumptions. Changing the market window can change the ranking, which is why HyperAgents puts the rules and the comparison beside the result.

Visitors can also draft a contender of their own. The builder combines a coin name, ticker, brief, preset, and simulated starting balance. A selected preset can be tested, and the draft can be exported as JSON. The brief helps express the identity; it does not create an autonomous AI trader or generate a new executable strategy from a prompt.

The next product direction is to bring coin creation into that experience through pump.fun, with HYPE as the intended quote asset. Pump's current supported-pair list includes a Solana HYPE representation supplied by Sunrise, from Wormhole Labs. That provides a documented pairing route for the concept; it is separate from the HYPE price feed used in the arena. [Pump supported pair assets](https://pump.fun/docs/custom-pairs)

Coin creation, creator-fee routing, autonomous AI execution, and live trading remain planned integrations. The present preview creates drafts and simulations. Pairing a future coin with HYPE would not itself connect a trading account, fund an agent, or entitle token holders to trading profits. Each of those capabilities needs its own concrete implementation and clearly stated behavior.

HyperAgents is an independent project inspired by the Hyperliquid ecosystem. Its dark forest and mint visual language gives the arena a coherent atmosphere, while its own symbol and agent identities make the project recognizable. It does not claim a partnership with Hyperliquid or pump.fun.

Different agents. One arena. Start with an approach, inspect its record, and decide what deserves to become a contender.