Roadmap

The TypeScript SDK. @paralock/sdk, @paralock/zcash, @paralock/adaptor, @paralock/evm and @paralock/lightwalletd on npm, MIT-licensed.

Why: any wallet, app or agent can add trustless ZEC-to-stablecoin swaps without a bridge or a custodian.

Both swap modes. htlc mode, and private mode with ECDSA adaptor signatures (experimental), on Tempo, Base, Ethereum, Arbitrum, Robinhood Chain and HyperEVM. See How it works.

Why: private mode leaves no hash or other shared value linking the two chains.

Escrows on six public testnets, with verified sources. ParalockHTLC and ParalockAdaptorEscrow on Tempo Moderato, Base Sepolia, Sepolia, Arbitrum Sepolia, Robinhood Chain Testnet and HyperEVM Testnet, each an exact bytecode match on Sourcify or Tempo's verifier. See deployments.

Why: anyone can check the code that holds the funds.

The first public testnet swaps. On 2026-10-12, Zcash testnet to Tempo Moderato: run 1 failed safely and refunded, and exposed a quote-expiry bug that the devnet's 3-second blocks hid; run 2, with the fix, swapped 0.015 TAZ for 19.39 OUSD in 3 min 39 s. Every transaction is in docs/TESTNET-RUNS.md.

Why: real block times and public infrastructure find bugs a local devnet can't.

The hardened protocol, in 0.3.0. The maker accepts and reserves before any ZEC moves (fundingExpiryHeight), its public swap view carries no adaptor data, and @paralock/zcash parses strictly and needs an explicit branch id and network.

Why: they close the bug from run 1 and other gaps found while hardening the code.

The protocol spec and test vectors. spec/PROTOCOL.md (protocol v1.1) with cross-language vectors in spec/vectors.

Why: a second implementation can prove it is byte-for-byte compatible.

The paralock CLI. npx paralock dev starts Zcash regtest, up to six EVM chains, the contracts and a maker, with no clone needed (0.3.0 is its first npm release). See CLI.

Why: a full swap on your machine in minutes, before touching a testnet.

Pay and payment links. Exact dollars to any address, from the CLI, the SDK or a link that opens the web app. See Payments.

Why: the person you pay never needs ZEC, gas or an app.

Docs for agents. A skill, llms.txt, every page as Markdown and a read-only MCP server. See For AI agents.

Why: coding agents integrate from the source of truth instead of guessing.

The Rust protocol core and EVM layer. paralock-core and paralock-evm, in the repository and checked against the same vectors.

Why: one core, checked once against the vectors, can serve every language instead of a port per platform.

The Rust SDK crate, paralock. Maker and taker engines, durable stores, the maker HTTP API (server and client), Zcash JSON-RPC and lightwalletd backends and a price feed, wire-compatible with @paralock/sdk.

Why: makers and services written in Rust can embed Paralock directly, and the bindings below build on it.

A Rust paralock binary. A native command line built on the Rust SDK.

Why: a maker can run as a single executable, without a Node.js runtime.

Publish the crates. paralock-core, paralock-evm and paralock on crates.io and docs.rs.

Why: until then, Rust users depend on the Git repository.

Bindings from the Rust core. napi for Node.js, wasm for browsers, uniffi for Swift and Kotlin, PyO3 for Python, and Go.

Why: wallets and apps on every platform get the same tested code instead of reimplementing the protocol.

A hosted testnet maker. Zcash testnet end to end against a maker that is always on.

Why: trying Paralock on a public network shouldn't require running your own maker.

A third-party security audit. Of the contracts, the protocol and the core.

Why: nothing should hold real funds on code that only its authors have reviewed.

Mainnet, after the audit. Starting with OUSD on Tempo.

Why: shielded ZEC holders get real dollars without a bridge, but only once the code has been audited.

A maker network. A request-for-quote network where many makers compete for each swap, with peer-to-peer offers on top; Paralock would earn a small routing fee.

Why: competing quotes give better prices and no single point of failure.

Dollars to ZEC. The reverse direction, so a taker can buy ZEC with stablecoins.

Why: today swaps only sell ZEC.

Refunds straight to a shielded address.

Why: today a refund lands on a transparent address and has to be shielded again.

Wallet integrations. "Pay in dollars" and "swap to dollars" inside Zcash wallets.

Why: most people will meet Paralock in a wallet they already use, not in a terminal.

MPP payments. A Zcash payment method for the Machine Payments Protocol.

Why: agents could pay HTTP 402 APIs straight from shielded ZEC, without buying stablecoins first.

Solana. An HTLC program with the same sha256 hashlock, and private mode through secp256k1_recover, which needs no cross-curve proofs.

Why: stablecoin liquidity on Solana, through the same protocol.

A fully shielded Zcash leg. Shielded notes controlled by a FROST 2-of-2 (RedPallas), Schnorr adaptor signatures, a cross-group DLEQ proof between Pallas and secp256k1, and a time-locked pre-signed refund. Open work: the cross-curve proof, Ironwood support in FROST, and wallet support for jointly controlled notes. See Privacy.

Why: the Zcash leg of a swap would become an ordinary shielded transaction, with only the EVM leg visible.