The agent-economy stack

Crypto infrastructure for AI agents

Autonomous agents are becoming first-class economic actors — they pay, prove, and coordinate across chains without a human in the loop. That takes four things crypto is uniquely good at. Cryptuon already ships production, MIT-licensed infrastructure for each of them.

Agentic payments (3) On-chain & verifiable AI (4) Real-world assets (2) Parallel EVM (1) Atomic composability & intents (3)

Agentic payments

3 products

Autonomous agents need to pay, escrow, and settle without a human in the loop. Cryptuon ships the on-chain rails — USDC escrow, sub-second settlement, and cross-chain coordination — that give agents a wallet that actually works.

Read the deep dive
NJ

Njord

Solana / Payments

Building

An on-chain affiliate marketing protocol on Solana — companies fund campaigns with USDC escrow and affiliates earn commissions that settle in roughly three seconds.

Anchor · SPL Solana
SA

Sarpoy

Solana / Game

Building

A Solana-native puzzle arena — creators deploy chat-bot puzzles backed by on-chain SOL prize pools, and solvers pay rising per-message fees on-chain to extract clues and claim the pot.

FastAPI · Anchor Solana
SB

Switchboard

Cross-chain

Building

A sub-400ms cross-chain state synchronization platform that uses Solana as a coordination layer, so one contract can read and write state across 50+ EVM and non-EVM chains without bespoke bridge code.

Solana · SDK Cross-chain

On-chain & verifiable AI

4 products

Decentralized inference, AI-drafted markets, and cryptographic proofs that a model — or a real-world outcome — is what it claims to be, with no trusted arbiter.

Read the deep dive
DF

DFPN

Solana / DePIN

Building

A decentralized coordination layer for deepfake detection on Solana — independent operators run detection models, a commit-reveal protocol stops collusion, and every verdict is anchored on-chain.

Rust · Anchor Solana
LM

SolanaLM

Solana / AI

Building

A hybrid decentralized AI network combining LLM inference and federated learning, with settlement and node incentives on Solana — operators earn SOL by serving inference and contributing GPU cycles.

Python · Solana Solana
ME

Mentat

Solana / Prediction markets

Building

AI-native prediction markets on Solana — markets are drafted by AI agents, curated by humans, and resolved by zkTLS proofs that cryptographically attest to real-world data, with no trusted arbiter.

FastAPI · zkTLS Trading
CR

commit-reveal

Crypto / Python library

Production

A pure-Python library for cryptographic commit-reveal schemes with optional Schnorr zero-knowledge proofs on secp256k1 — zero external dependencies, MIT licensed.

Python · PyPI Libraries

Real-world assets

2 products

Institutional-grade execution and verifiable data provenance for tokenized real-world assets — whale-sized trades without market impact, and compliant data sharing with immutable audit trails.

Read the deep dive
DM

DataMgmt Node

Enterprise data

Building

A decentralized enterprise data-management platform combining EVM-compatible blockchains, a Kademlia DHT P2P network, and end-to-end encryption for secure, compliant data sharing.

Python · EVM Cross-chain
MM

Moby Market

DeFi / Institutional

Building

Institutional-grade trading infrastructure for whale-sized DeFi trades on Solana — an open-source Rust suite delivering OTC, TWAP/VWAP, zero-knowledge privacy, and intent-based execution.

Rust · Anchor Trading

Parallel EVM

1 product

A throughput ceiling is a business risk. Cryptuon builds an embeddable EVM with wave-based parallel execution so agent-driven, high-frequency workloads have room to run.

Read the deep dive
ZE

Zig-EVM

EVM / Runtime

Production

A high-performance, embeddable Ethereum Virtual Machine in Zig — 96+ opcodes, wave-based parallel execution for a 5-6x throughput gain, and native FFI bindings for Python, Rust, JavaScript, and C.

Zig EVM

Atomic composability & intents

3 products

Agents operate across many chains at once. All-or-nothing cross-rollup swaps, cross-chain state sync, and commit-reveal fairness turn a fragmented multi-chain stack into one coordinated execution surface.

Read the deep dive
TE

Tesseract

Cross-chain / DeFi

Production

A production-ready cross-rollup atomic swap protocol for Ethereum L2s — Vyper contracts plus a Rust relayer enforce all-or-nothing swaps with commit-reveal MEV protection, no bridges or wrapped assets.

Vyper · Rust Cross-chain
SB

Switchboard

Cross-chain

Building

A sub-400ms cross-chain state synchronization platform that uses Solana as a coordination layer, so one contract can read and write state across 50+ EVM and non-EVM chains without bespoke bridge code.

Solana · SDK Cross-chain
CR

commit-reveal

Crypto / Python library

Production

A pure-Python library for cryptographic commit-reveal schemes with optional Schnorr zero-knowledge proofs on secp256k1 — zero external dependencies, MIT licensed.

Python · PyPI Libraries

Frequently asked questions

What is the "agent economy" in crypto?

The agent economy is the emerging pattern where autonomous AI agents hold wallets, pay for services, and settle transactions on-chain without a human approving each step. It needs three things crypto is uniquely good at: programmable payments, verifiable computation, and cross-chain coordination — exactly the primitives Cryptuon builds.

Can AI agents hold and use crypto wallets?

Yes. An AI agent can control a wallet, deposit stablecoins, and pay for APIs or data programmatically. The hard parts are settlement speed, spend limits, and trust — an agent should not be able to drain a balance or act on unverified data. Cryptuon's njord (USDC-escrow, ~3-second settlement) and mentat (zkTLS-verified resolution) address the payment and trust legs respectively.

What is x402 / HTTP-402 for agent payments?

HTTP status code 402 ("Payment Required") is being revived as a standard where a server responds to a request with a payment challenge, and the client (often an AI agent) answers with a signed on-chain payment before the service executes. It turns any API call into a metered, pay-per-use transaction — a natural fit for agents that pay as they go.

How do you verify an AI model ran correctly on-chain?

You cannot re-run a large model inside a smart contract, so the pattern is to attach a proof to the result: a zkTLS attestation of the source data, a commit-reveal scheme that stops operators from copying each other, or a decentralized network where multiple operators run the same model and outliers are slashed. Cryptuon uses these across solanalm, mentat, dfpn, and the commit-reveal library.

Why do AI agents need cross-chain infrastructure?

Agents optimise across whatever chain has the best price, liquidity, or data — which means they operate on many chains at once. Without atomic guarantees, a multi-chain action can half-complete and strand funds. Tesseract enforces all-or-nothing cross-rollup swaps and Switchboard synchronises state across 50+ chains so an agent sees one coherent execution surface.

Does Cryptuon build the AI agents themselves?

No. Cryptuon builds the open-source, MIT-licensed infrastructure the agent economy runs on — payment rails, verifiable-AI primitives, an embeddable parallel EVM, and cross-chain coordination. Teams building agents plug these in rather than reinventing settlement, proofs, and interoperability.

Building an agent that needs to transact?

Plug into open-source rails for payments, verifiable AI, and cross-chain coordination instead of reinventing settlement, proofs, and interoperability.