Brevis primary-source snapshot

Fetched: 2026-05-08 UTC

1) Official guide: “The Infinite Compute Layer for Everything”

Source: https://blog.brevis.network/2025/10/17/the-brevis-guide-the-infinite-compute-layer-for-everything/

Selected excerpts from the guide:

Instead of every validator re-executing computation, one prover executes it off-chain and generates a mathematical proof that the computation was done correctly.

This “compute off-chain, verify on-chain” paradigm gives blockchain applications unlimited computational freedom without sacrificing the trustlessness that makes decentralization valuable.

The ZK ecosystem reflects this reality. Specialized teams have built proving infrastructure optimized for specific workload types.

Additional signals from the guide:

  • Brevis is explicitly arguing that proving workloads are heterogeneous and should not be served by one vertically integrated proving stack
  • The guide ties three product surfaces together: Pico zkVM, the ZK Data Coprocessor, and ProverNet
  • The important control surface is not just proof generation; it is who coordinates hardware specialization, latency classes, and access to proving capacity

2) ProverNet launch post

Source: https://blog.brevis.network/2025/12/08/brevis-provernet-mainnet-beta-is-live/

Selected excerpts from the launch post:

On November 17th we released the ProverNet whitepaper, detailing our design for a decentralized marketplace where any application can access ZK proving capacity and any prover can compete to serve it.

The marketplace is fully operational. Proof auctions run continuously, matching requests to provers and clearing payments in real-time.

From this point on, applications get access to specialized capacity without vendor lock-in while provers can now find workloads matched to their hardware.

Additional signals from the launch post:

  • The mainnet-beta post is the clearest first-party statement that Brevis now treats the marketplace itself as a core product, not just internal infrastructure
  • The post also makes the workload-diversity argument explicit: many partners and proof types produced the need for a shared market rather than bespoke proving setups
  • The post says USDC is used during beta with BREV intended as the full-mainnet settlement asset later, which matters for reading the token and market design

3) ProverNet docs repo README

Source: https://raw.githubusercontent.com/brevis-network/provenet-docs/main/README.md

Selected excerpts from the README:

ProverNet operates as a decentralized marketplace that bridges the gap between applications requiring zero-knowledge proofs and the specialized hardware capable of generating them.

The BREV token serves as the core of the ProverNet economy, aligning incentives and facilitating value exchange across the network.

Additional signals from the README:

  • The README makes the participant model explicit: developers, users, provers, and stakers all have distinct roles in the market
  • It also states that BREV is used for payments, economic security, reward distribution, and governance, which clarifies that ProverNet is intended as a full proving economy rather than a simple job board

4) Prover operations repo README

Source: https://raw.githubusercontent.com/brevis-network/prover-network-ops/main/README.md

Selected excerpts from the README:

Docs and utilities for operating within the Brevis Network prover marketplace, with an emphasis on the step-by-step manual for standing up a prover node.

request-proof – Submit proof requests to BrevisMarket.

Additional signals from the README:

  • The public ops repo is useful evidence that ProverNet is operable infrastructure rather than only a whitepaper concept
  • The repo shows that Brevis exposes concrete operator and requester workflows through CLI tooling, configs, and walkthroughs

5) Prover node manual

Source: https://raw.githubusercontent.com/brevis-network/prover-network-ops/main/docs/run_prover_node.md

Selected excerpts from the manual:

A prover node must run both the Pico proving service (to generate proofs) and the bidder service (to interact with the proving network).

BREV is the staking token on Base mainnet.

Additional signals from the manual:

  • The manual documents a two-part operating model: proving service plus bidder service
  • It exposes meaningful implementation detail about staking, commission rates, prover versus submitter accounts, supported hardware paths, and Base-mainnet contract touchpoints
  • This is strong evidence that the proving market has real operational boundaries and hidden authority surfaces around bidder logic, fee ceilings, whitelist/blacklist filters, and commission policy

6) Proof-request guide

Source: https://raw.githubusercontent.com/brevis-network/prover-network-ops/main/docs/request_proof.md

Selected excerpts from the guide:

The easiest way to submit requests is via the CLI utility, which batches multiple entries and handles encoding. On Base mainnet, you must hold sufficient BREV to cover the proof fee.

Call requestProof on BrevisMarket with the parameters gathered above.

Additional signals from the guide:

  • The request flow makes the market mechanism legible: proofs are parameterized by verification key, public-values digest, fee ceiling, minimum prover stake, and deadline
  • This means apps do not simply “buy proving”; they specify deadline, security floor, and maximum price, which turns proof generation into a programmable market primitive

7) Whitepaper status

Source: https://brevis.network/whitepaper/provernet.pdf

Notes:

  • A canonical ProverNet whitepaper exists at the official URL above
  • During this pass, the blog + docs + ops materials were easier to extract and more operationally concrete than the PDF itself, so they were used as the main evidence surface for summary writing

Notes

  • Brevis is best read as universal proving infrastructure plus a proving-market coordination layer, not only as a zk coprocessor product.
  • The strongest reusable comparison is against projects that separate proving generation, proving verification, and proving-market coordination into different layers.
  • Good comparison partners for later work: Lagrange, Axiom/OpenVM, RISC Zero/Boundless, and zkVerify.