# Airbender * ZKsync Airbender is a new ZK proof system. * [**ADI Network**](/adi-network-components/overview.md) utilized an **“operating system”** based on ZKsync OS**.** ### What is Airbender? Airbender is ZKsync's next-generation proof system, purpose-built to enable efficient ZK proofs of RISC-V bytecode execution. Built on the foundation of highly optimized STARK/FRI implementations, Airbender is designed to support long-term scaling strategy by being fast, cheap, and flexible to a wide range of use cases (without compromising security). ### Key Features * RISC-V 32I+M instruction set * AIR constraints * Optimized DEEP STARK/FRI proofs * Mersenne31 prime field * Blake2s + Blake3 hash function The name “Airbender” is a reference to AIR constraints (Arithmetic Intermediate Representation). ### Proving Pipeline The proving process follows a structured six-stage pipeline designed for optimal performance and resource utilization: * **Stage 1 - Witness Generation**: Computes Low-Degree Extensions (LDEs) of witness data and generates trace commitments. This stage establishes the foundational data structures for the proof. * **Stage 2 - Lookup and Memory Argument**: Configures lookup tables and memory arguments, establishing the infrastructure for efficient verification of memory operations and table lookups. * **Stage 3 - STARK Quotient Polynomial**: Computes the primary STARK quotient polynomial, which encodes the constraint satisfaction proof for the main circuit. * **Stage 4 - DEEP Polynomial**: Implements FRI batching optimizations to reduce the overall proof size and verification complexity. * **Stage 5 - FRI IOPP Generation**: Produces the final Interactive Oracle Proof of Proximity (FRI proof) that enables efficient verification. * **Stage 6 - SNARK Wrapper:** Wrap the final FRI proof into a final FFLONK proof, which gets posted and verified on-chain. --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.adi.foundation/adi-network-components/airbender.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.