Torque

What Torque is

Torque is a decentralized perpetuals exchange. It runs as a builder-code frontend on Hyperliquid, so every order settles on Hyperliquid's on-chain order book while Torque handles the experience, the markets, and the incentives around it.

Two things make Torque different from the venues people already use. First, it opens markets that nobody else lists: pre-launch tokens, AI-agent tokens, memecoins, and crypto-native event markets, alongside the majors. Second, it hands most of the value back to the people who actually trade. Sixty percent of the token supply is reserved for the community, and the buyback is funded from real trading fees rather than a treasury raise.

Torque is bootstrapped. There was no large private round and no venture treasury behind it. That constraint shaped the design. If the protocol is going to reward its users at the scale we want, the money has to come from usage, not from selling tokens to insiders first.

The problem we started with

Perp trading has consolidated around a handful of venues that all list the same thirty or forty assets. BTC, ETH, SOL, a rotating set of large caps. The fee race on those markets has already run its course. Several venues now charge close to nothing on majors, and a maker can trade for a basis point or less. Competing there on price alone is a losing game for anyone without a balance sheet to subsidize it.

Meanwhile, the assets people are most curious about are the ones you cannot get exposure to. A token that has not launched yet. An AI agent that is trending this week. A memecoin with real volume and no perp market. The outcome of an event that the whole timeline is arguing about. These markets are either impossible to trade or locked inside closed OTC desks with wide spreads and no transparency.

That gap is the opening. Torque is built for the tail, not the head.

How it works

Torque signs and routes orders through Hyperliquid using builder codes. When you place a trade on a major, it goes to Hyperliquid's real order book and fills against real liquidity. You keep the depth and execution of one of the deepest on-chain venues in the market, and you get Torque's markets and rewards on top.

Trading is non-custodial. Torque never holds your funds and never takes custody of your account. To sign orders without a wallet popup on every trade, the app uses a session agent key that is approved once. That key can place orders on your behalf and nothing else. It cannot withdraw, transfer, or move funds out of your account. We ran a review of the codebase specifically for this and found no path that touches user funds beyond placing trades you authorize.

Market data is served through the protocol rather than pulled directly in the browser, which keeps prices and books working on mobile networks and behind firewalls where a direct connection would fail. Pre-launch markets are priced from live on-chain OTC data, so a pre-market shows a real price when one exists and an honest "opening soon" when it does not. There are no invented numbers anywhere in the app.

Markets

Torque lists two kinds of market.

The majors are BTC, ETH, SOL, HYPE, and the other large caps that route to Hyperliquid. These give traders deep liquidity and tight spreads, and they are where most day-to-day volume lives.

The long-tail is where Torque is the venue rather than a frontend on someone else's book. This is the part that does not exist anywhere else in a clean, on-chain form:

• Pre-launch token markets, priced from live OTC data, so you can take a position on a token before it lists.
• AI-agent tokens and indices, a category that is growing faster than any listing process can keep up with.
• Memecoins with real volume that the large venues will not touch.
• Crypto-native event markets tied to outcomes the community actually cares about.

We deliberately stay out of stocks, foreign exchange, commodities, and pre-IPO equity. Those markets are crowded with incumbents who do them better, and they carry regulatory weight that does not fit a permissionless protocol. Torque plays where it can win.

Fees

The fee model follows the market structure. On majors, where traders can go anywhere and price is the only lever, Torque keeps fees low and competitive. There is no point charging a premium on BTC when the venue next door is close to free.

The premium lives where Torque is the only place to trade. In the long-tail markets, where the protocol lists an asset that has no other on-chain home, a higher fee is fair and the market bears it. Traders are paying for access to something they cannot get elsewhere, not for a commodity.

A share of fees funds the buyback. Another share covers the cost of running the protocol. The point of the split is simple: the more the protocol is used, the more value flows back to the token and the people holding it, without the team needing to sell tokens to keep the lights on.

Points and seasons

Torque runs its early growth through seasons. Each season has its own points program, and points reward the behavior that makes an exchange healthy rather than the behavior that games it.

Points go to real trading activity, to holding positions over time instead of flipping in and out, and to referrals that bring in genuine traders. The program has safeguards against wash trading and sybil farming, and the exact weighting is kept private and adjusts over time so that it cannot be reverse-engineered and gamed. We will not publish the formula, for the same reason no serious rewards program does.

Season 0 is the testnet genesis season, live now. It sits on Hyperliquid testnet with no real money at risk, so anyone can connect a wallet, get funded from the in-app faucet, and start trading. Season 0 does not show or award points. Your activity, your volume, your holding, and the real traders you refer, is tracked from your first trade and ranks you on a live activity leaderboard. Your Season 0 reward is granted retroactively when mainnet opens, after a review that filters out farming. Testnet funds are free, so paying points out there would only pay sybils to farm, which is why the reward is decided later from real behavior rather than a number on a screen.

The real distribution weight sits in Season 1, which opens on mainnet and runs for forty-four weeks. It brings real trading, real liquidity, and points that reward genuine activity: trading, holding positions through time, and referring real traders. Both seasons count toward the token distribution at the end of Season 1, with mainnet carrying the majority. Showing up honestly in Season 0 is how you get in early. Being active on mainnet in Season 1 is where the weight is.

We are not publishing how large the airdrop will be or how it converts. That is deliberate. Announcing a number early invites farming and locks in a design before the data is in. What we will say is that the community allocation is real, it is the majority of supply, and the people who show up early and trade honestly are the ones it is built for.

The TRQ token

TRQ is the token for the protocol and the app. It shares one name with the product on purpose. There is one brand, one asset, one thing to hold.

Sixty percent of the total supply is reserved for the community. That is the number that matters, and it is fixed. The rest covers liquidity, protocol-owned reserves, and the cost of building and running Torque over the long run. We are not going to pretend a project runs on zero, but the balance is set so that the people using the protocol own most of it.

Value flows back to the token through a buyback funded by trading fees. When the protocol earns, it buys TRQ. A spot listing feature lets projects pay through a Dutch auction for the right to deploy a spot market on Torque, and the proceeds from those auctions are split between buying and burning TRQ and the fee waterfall. Every one of these mechanisms is funded by real activity, not emissions.

Holding TRQ also pays. A share of the profit the business generates is distributed to token holders every quarter. This is a cash dividend paid from the protocol's actual earnings, not new tokens minted to fake a yield. When Torque does well, holders are paid on a fixed schedule, and the stronger the quarter, the larger the distribution. It turns the token into a claim on a working business rather than a bet on price alone.

The token generation event fires at the end of Season 1, and it is gated on the protocol hitting real metrics rather than an arbitrary date. Volume, retention, and protocol-owned liquidity have to clear a bar before the token goes live. A token launched into thin usage is a token that dumps. We would rather earn the launch than schedule it.

Roadmap

Season 0, now. Testnet genesis. The app is live on testnet and runs for four weeks. Public trading, the points engine, referrals, and faucet onboarding are all running. The goal here is to seed a real community, harden the system under real use, and prove the whole loop from trade to points to leaderboard before any money is on the line.

Season 1. Mainnet. Season 1 opens on mainnet and runs for forty-four weeks. Real fees, real trading, real buyback. The full points model comes online with the mainnet fee economics behind it. This is the season that earns the token generation event, gated on the metrics above. Referral rewards run on real fee flow, and the buyback starts turning usage into token demand.

Season 2. Post-token. The token is live. Season 2 shifts from a points program to token-based rewards and opens up governance over the parameters that were held private during bootstrapping. Emissions are tied to protocol growth so that new traders add to the pool rather than only diluting it.

Beyond. Two directions open up once volume and capital justify them. First, graduating from a builder-code frontend to deploying Torque's own perp markets directly, which gives the protocol control over listings and margin rather than depending entirely on the underlying venue. Second, expanding the long-tail: more pre-launch coverage, deeper AI-agent and event markets, and creator markets. The thesis stays the same the whole way. Torque is the place to trade the markets that do not exist yet.

Where this is going

The venues that matter in five years will not be the ones with the lowest fee on BTC. That race is already over. They will be the ones that made new markets tradable first, and that returned enough of the upside to their users that the users stuck around.

Torque is a bet on both. Open the tail, price it fairly because you are the only venue there, and hand the majority of the token to the people generating the volume. If that works, the flywheel runs on its own: more markets bring more traders, more traders generate more fees, more fees fund the buyback and the rewards, and the rewards bring more traders back.

We are building it in the open, starting on testnet, with the community holding most of what gets created. That is the whole idea.