Background

During the past 1.5 years, We (Manta.Network) developed the on-chain privacy solution for DOT/SAMA eco-systems using zkSNARKs, here are some highlights:

Dolphin Testnet V1(unincentivized)

• 60,000 + transactions
• 6,000+ distinct addresses
• send more private transaction powered by zkSNARK in 3 months than Tornado.cash in 1 year

"The Dolphin Testnet] is really good... The experience is really pleasing, congrats to
the team.”
(Thibaut Sarda, Core Developer, Polkadot.js)

Dolphin Testnet V2

• 1/10 ZCash's (Sapling) constraints count in zero-knowledge-proof circuit
• Improved prover performance from 30 sec to 2 sec

As the first project in the substrate space (and very few in the entire Web 3 space) working on zero-knowledge-proof based developed, we experienced a lot of difficulties in terms of ZKP tooling. As a result, we developed a lot of internal tooling that could benefit peer substrate teams deploying zero-knowledge-proof techniques securely and quickly into their project.

Potential Applications

• DID projects like KILT can leverage OpenZL to make their KYC/AML process more decentralized
• Cross-chain Bridges can leverage OpenZL to make the bridges high-performance and trust-less
• DAO tooling project can leverage OpenZL to build MACI for governance
• Project like Astar can leverage OpenZL to build ZKRollups

and many more

OpenZL Design

OpenZL is an open-source library that helps practioners (especially in Web3 space) to develop and deploy secure, high performance zero-knowledge proof code in production. It tries to bridge the gap between low level cryptographic primitives and devlopers' need to build scalable protocols using zero-knowlege proof cryptography securely and quickly. More specifically, many developers today want to leverage zero-knowledge proof systems to build powerful protocols like ZCash/Manta/ZKSync. However, they are facing two less than ideal choices; first, building a protocol using high-level languages like Circom or Cairo loses many performance optimization opportunities, and second, building the protocol directly using libraries like arkworks/groth16, zk-garage/plonk, or microsoft/nova requires expertise in cryptography and can be very error-prone. Also, zero-knowledge proof systems are a moving target. There have been many new, and "better", proof systems coming out every 2-3 years (BCTV -> Groth16 -> Plonk -> Nova). OpenZL tries to solve this problem by building flexible, proof-system agnostic, and extensible libraries for Web3 practitioners.

OpenZL consists of 3 parts:

• Gadget libraries: a library of gadgets that developers can use as building blocks for their protocols. The initial range of the gadgets includes accumulators (merkle tree with zero-knowledge membership proof), zk-friendly hash functions (poseidon hash), and commitment schemes. The gadget libraries are programmed in eclair.
• Embedded Circuit Language And Intermediate Representation (eclair): An embedded DSL in Rust that describes circuit logic. eclair leverages Rust's expressive type system to rule out certain classes of errors during the circuit construction.
• Adaptors to Proof Systems: Adaptors that convert circuit logic in eclair to the constraint systems used in different proof systems. The initial supported proof systems are arkworks/groth16, zk-garage/plonk, and microsoft/nova.

Technical Details and Budget:

Please find the technical details and budget in the following page:
https://github.com/OpenZkLib/openzl/blob/main/proposal.md

Would love to get constructive feedbacks.