The missing corporate-actions layer for tokenized stocks. On-chain dividends, stock splits, and record-date semantics — for the tokens that already exist.
Parvalon is the missing corporate-actions layer for tokenized stocks — on-chain dividends, splits, and record-date semantics for the tokens that already exist. It is a permissionless overlay: it works on tokenized stocks we do not control, with no token changes and no issuer integration.
The problem. Tokenization solved issuance and trading for ~2,000 tokenized stocks, but not what happens after a share is minted — dividends, splits, record dates. A holder of a tokenized share has no on-chain rail to receive a dividend; DeFi has no trustworthy signal when a stock splits or goes ex-dividend (a systemic risk once RWAs are used as collateral); and AI agents have no machine-readable feed to react to.
How it works.
Announce a corporate action on-chain with correct record-date semantics.
Snapshot every holder's balance at the record block from public Transfer logs, build a Merkle tree, and publish the root — anyone can independently reproduce it.
Fund & claim pro-rata dividends in USDG against that root: O(1) gas per claim, claim-on-behalf (relayable / gasless), funds always settle to the holder.
Consume a standardized event stream (our draft CAE-1 EIP) plus a REST feed, so lending markets, AMMs and agents can finally react.
Why it's different. Paying dividends on-chain normally requires owning the token (transfer hooks, rebasing, issuer integration). Parvalon inverts that: it reconstructs holders of any standard ERC-20 via eth_getLogs, so it runs on the real Robinhood TSLA / AMZN / PLTR / NFLX / AMD tokens today with nothing required from the issuer. Two immutable, non-upgradeable contracts — a registry that governs state and a distributor that custodies and settles value; record dates are enforced on the true L2 block via the Arbitrum ArbSys precompile.
Proven live on Robinhood Chain testnet: all five real tokenized stocks have a funded, CLAIMABLE dividend on-chain — ~917,000 holders snapshotted, ≈13.6M USDG funded, with real claims settled on-chain. Engineering: 84 Foundry tests (fuzz + invariants), Slither 0 high / 0 medium, a parallel snapshot scanner over ~3M Transfer logs, and a server-side proofs API that serves per-holder proofs at 184k-holder scale.
<p>Built for the buildathon as a Solidity + Next.js/TypeScript monorepo. What we shipped:</p><ul><li><p><strong>Protocol (2 immutable contracts).</strong> <code data-inline="true" spellcheck="false">CorporateActionRegistry</code> + <code data-inline="true" spellcheck="false">DividendDistributor</code>: the announce → publish-root → fund → claim lifecycle, Merkle verification + a claim bitmap, per-action solvency isolation, and claim-on-behalf. <strong>84 Foundry tests</strong> (unit + fuzz + invariants), <strong>Slither 0 high / 0 medium</strong>.</p></li><li><p><strong>Arbitrum/Orbit record-date fix.</strong> Record dates are enforced on the L2 block via the <code data-inline="true" spellcheck="false">ArbSys</code> precompile (raw <code data-inline="true" spellcheck="false">block.number</code> on Orbit is the L1 height), so the on-chain guard and the off-chain snapshot share one clock.</p></li><li><p><strong>Live on Robinhood Chain (46630).</strong> Deployed the registry, distributor and oracle source, then gave <strong>all 5 real stocks a funded, CLAIMABLE dividend</strong> — ~917,000 holders snapshotted, ≈13.6M USDG funded — with real on-chain claims settled.</p></li><li><p><strong>Snapshot at scale.</strong> A parallel <code data-inline="true" spellcheck="false">eth_getLogs</code> scanner that reconstructs ~184k holders per token from ~3M Transfer logs in ~10 minutes, producing deterministic, independently re-derivable Merkle roots.</p></li><li><p><strong>Proofs API.</strong> A server-side per-holder proof lookup (sharded store) so a 264 MB / 184k-holder proof set never ships to the browser — claiming works at full scale.</p></li><li><p><strong>dApp (Next.js).</strong> <code data-inline="true" spellcheck="false">/claim</code> (a table showing each holder's snapshot balance → claimable dividend, claim-on-behalf, explorer-linked receipts), plus <code data-inline="true" spellcheck="false">/issuer</code> console, <code data-inline="true" spellcheck="false">/feed</code>, <code data-inline="true" spellcheck="false">/faucet</code>, and in-app <code data-inline="true" spellcheck="false">/docs</code>.</p></li><li><p><strong>Ecosystem tooling.</strong> A typed TypeScript SDK, a deterministic snapshot CLI, a solvency monitor, a draft <strong>CAE-1</strong> EIP (Corporate Action Events standard) + a <code data-inline="true" spellcheck="false">/api/actions</code> feed, and a runnable example agent that claims-on-behalf from CAE-1 events.</p></li><li><p><strong>Docs.</strong> Architecture, threat model, a frozen integration contract, an honest limitations doc, and a P0/P1/P2 production-readiness roadmap.</p></li></ul>
<p><strong>Pre-seed / bootstrapped — no external funding raised to date.</strong> Parvalon was built independently for the Arbitrum Open House London Buildathon. We have taken no investment and have no token.</p><p><strong>Seeking:</strong> the Arbitrum Foundation milestone grant and Robinhood Chain ecosystem support to fund the production path — an external security audit, governance hardening (Timelock + multisig handover, already coded), a payout-token allowlist, and a pilot with a tokenization platform / issuer.</p>
