Products

Market views

You don't pick from a menu of tokens. You pick a view, and each view is a deterministic strategy with rules you can read.

A normal perp DEX asks: market, long or short, leverage, size. PerpTokens asks: market, view, direction, intensity. One underlying Phoenix market can back many strategy tokens (trend long, trend short, range long, range short). These are different ways to express a SOL view, not thirty confusing SOL markets. Three views are live today: Trend, Range, and Pair.

Trend: directional conviction

A trend vault targets constant leverage against mark-to-market equity. When price moves in its favor, equity grows, the target notional grows with it, and the vault adds exposure into the move. In a clean, persistent trend this compounds and can outperform a static perp position.

The flip side is mechanical: maintaining target leverage means buying after price rises and selling after it falls. In a choppy, mean-reverting market that is buy-high/sell-low behavior, and rebalancing costs eat returns.

Trend tokens are for directional conviction. They work best when the move persists; they struggle when the market trades sideways.

Range: range-bound markets

A range vault flips the objective. Instead of targeting notional against mark-to-market equity, it targets notional against backing collateral. When price rallies, notional runs high relative to backing and the vault sells exposure; when price dips, notional runs low and the vault buys. That produces sell-high/buy-low behavior without any admin-set price band: the reference point is the vault's own balance sheet, not a manually chosen anchor.

True mark-to-market leverage still acts as the safety rail: every exposure-increasing move is checked against a hard leverage cap, so the strategy can't buy itself into the danger zone while fading a move.

Both biases exist: a long range token stays structurally long and harvests the swings around that exposure; a short range token does the same on the short side.

Range tokens are for range-bound or mean-reverting markets. They work best when the market oscillates; they struggle when it trends hard and the token keeps fading a move that continues.

Pair: relative value, long one and short another

A pair vault runs a multi-leg long/short book inside a single token: long one market and short another (for example, long SOL / short ETH) so the token expresses a relative-value view rather than a directional one. It pays off when the long leg outperforms the short leg, regardless of which way the broader market moves.

Each pair vault can span up to six Phoenix markets, and every leg runs on isolated margin. A liquidation on one leg stays contained to that leg instead of cascading across the whole position, so a blowup on one side doesn't take the rest of the book with it.

Pair tokens are for relative-value conviction. They work best when your spread thesis plays out; they struggle when the two legs move together and the spread doesn't.

What's next

The same primitive supports a broader catalog of deterministic strategies: funding carry, hedges, reserve-backed upside, and more. The catalog grows deliberately. Each design has to be easy to explain, easy to simulate, and proven before the next one ships. The protocol is not in the black-box strategy manager business.

Regimes cut both ways

Every token is designed for a market regime, and the wrong regime loses money: trend tokens bleed in a range, range tokens bleed in strong trends. Choose by your view, not by what performed last week. See Risks.