C-VIX 30D index
A constant 30-day volatility index based on BTC and ETH options. PUSHIN currently displays generated practice values.
PlannedThe complete sequence: platform foundations, operational hardening, testnet, contract upgrades, terminal and API, security, mainnet, token utility and research.
Current release: browser-based practice trading. The infrastructure and on-chain stages below are plans for PUSHIN.
A constant 30-day volatility index based on BTC and ETH options. PUSHIN currently displays generated practice values.
PlannedNormalise Binance, Hyperliquid and Lighter rates to an eight-hour basis. Live collection is planned.
PlannedBrowser-local market and limit orders, virtual balances, positions and trade history.
Available locallyA scoped authorisation design for a future relayer. Local practice accounts do not request signatures.
PlannedSpecify the payout vault, oracle dampener, skew fees and settlement invariants.
PlannedPlanned keeper, settler, indexer and liquidator services.
PlannedDeploy and verify PUSHIN contracts before any public on-chain trading.
PlannedThe frontend is hosted on Vercel. A relayer and persistent database remain to be integrated.
PlannedToken deployment, supply and distribution must be confirmed before a launch.
PlannedUptime checks on the health endpoint, the relayer's heartbeat and the keeper's and relayer's gas balances, paging a person instead of writing a log line.
PlannedExceptions from the web app and the relayer collected within one system, each with the request or round that caused it.
PlannedTypecheck, lint, the core and contract suites, ABI drift, Slither and a production build on each push, and the API end-to-end suite against each preview. At present, each check is run by hand.
PlannedPer-address and per-IP limits on orders, sessions, nonces and collateral moves, so an open venue cannot be flooded.
PlannedA lease in Postgres that the relayer must hold to write, so a second instance waits as a standby instead of sending each keeper transaction twice. At present, it is a convention.
PlannedReplace 90 days of seeded FR-BASIS rows with the venues' own published funding history. The backfill is written; it hasn’t been run on production.
PlannedLive collection began on 20 September 2026 and covers 90 days on 19 December. That day the seeded rows are deleted, and nothing else is.
PlannedCheck each formula, rule and status line on the docs and protocol pages against the contracts, and rewrite what predates them.
PlannedMove all four contracts to a Safe: transferOwnership from the deployer, then one Safe batch of acceptOwnership. At present, the deployer key owns them.
PlannedA second deployment, such as testnet.the practice venue, with its own relayer and database on the chain venue, so the chain venue runs in public without moving the main site off paper.
PlannedSeparate keeper, settler and liquidator keys, each funded for one job and watched for its balance, so a leaked key can do only that job.
PlannedRead isSolvent, isBacked and the pool's free liquidity each block, alert on any breach, and chart them on the risk dashboard.
PlannedAnother option surface in the C-VIX aggregate, so the index, and each C-VIX position with it, doesn’t stop when Deribit does.
PlannedExercise liquidation on testnet. It is proven so far only on a local chain, where the test can move the index itself.
PlannedTop up an open position to move its liquidation level away. At present, the only way to take risk off a position is to close it.
PlannedClose part of a position, with its margin, escrow and accrued costs split pro rata. At present, each order opens a new position and a close takes all of it.
PlannedCloses signed in advance with a trigger level. The contract checks the trigger against the oracle, so the relayer cannot fire one early.
PlannedMarkets as configuration, each with its kind, leverage cap, feed and limits, instead of three constants compiled into the contracts. A listing then needs no redeployment.
PlannedLimits per market and per account. At present, the only bound is the pool's free liquidity.
PlannedThe oracle settles on the median of values signed by several independent keepers, so a single compromised keeper cannot move it. At present, any allowed keeper can push within the band.
PlannedIf a feed stays stale past a long, fixed bound, anyone can settle that market's positions at its last fresh level. Today a stopped keeper freezes exits until it returns.
PlannedChanges to the borrow rate, depth, liquidation bounty, oracle age and spread step wait out a public delay. Pauses stay immediate, since they only protect.
PlannedThe fee split on the tokenomics page, applied by a contract. At present, each fee goes to the payout pool.
PlannedThe paper matcher gains each v2 feature, and a differential test runs one order sequence through both venues and compares each balance.
PlannedAdd margin, partial close, reduce-only, stop-loss and take-profit, from the order ticket and the positions table.
PlannedA notification when a position nears its liquidation level, and when an order fills or expires.
PlannedFills, closes and each ledger row as CSV, for bookkeeping.
PlannedDocumented REST and streaming endpoints and a typed client built on @prismperp/core, for bots and market makers trading through session keys.
PlannedAn independent firm reviews the tagged commit. Each finding is fixed with a regression test, and the report is published naming the commit it covers.
PlannedThe four invariants held over millions of random calls instead of 128,000, and checked again by a second fuzzing engine.
PlannedA public bounty with rewards scaled to the funds at risk, live ahead of the first real deposit.
PlannedKeeper and settler keys held by a key-management service that signs without exposing them, not in environment variables, with a rotation runbook.
PlannedWho can pause what, how fast, and what is posted on X, rehearsed on testnet with the Safe's real signers.
PlannedThe audited bytecode on Robinhood Chain (4663) with Paxos's USDG as collateral, owned by the Safe through the timelock starting with the first block, and verified on the explorer.
PlannedThe payout pool funded with real USDG. Each position reserves eight times its margin from it, so the pool can back open margin up to an eighth of its size.
PlannedLow open-interest and deposit caps at first, raised on a published schedule as the pool grows and the book stays two-sided.
PlannedFeed freshness, keeper health, pool solvency and any pause, read from the chain and shown to everyone.
Plannedthe practice venue moves to the chain venue, and the paper venue stays open for practice.
PlannedStake $PUSHIN for the share of trading fees the tokenomics page describes, paid in USDG. Without token emissions: rewards come from fees, not new tokens.
PlannedThe fee router's buyback share spent in the $PUSHIN pool on a public schedule, with each purchase and burn visible on-chain.
PlannedVotes on oracle venue weights, dampener parameters and new listings, carried out by the Safe through the timelock at first and on-chain later.
PlannedThe keeper share of fees paying for the gas that keeps the oracle fresh and orders settling. There is no token treasury to draw on.
PlannedReserve the book's worst case instead of the sum of each position's cap. Longs and shorts on one index largely offset, so the same pool could back more open interest, once the bound is proven.
ResearchA vault with shares, so others can fund the payout pool and take its side of the book. Who carries that risk is a design decision before it is a technical one.
ResearchFurther volatility tenors and underlyings where an option surface is deep enough to replicate, and further basis pairs as the native venue lists them.
ResearchOne margin account across positions instead of isolated margin for each.
ResearchAnyone may run a keeper, bonded and paid per fresh push, so the oracle doesn’t depend on one operator.
Research