Native, heterogeneous compute

Ritual’s thesis is rooted in the defining hypothesis that the complexity of what you can do on blockchain networks will dramatically grow over time.

To that extent, Ritual is focused on being the most expressive blockchain in existence, hyperscaling the expressive compute that can be performed on a blockchain.

Ritual firmly believes that developers want to consume this compute via the tools they are already familiar with, beginning by supporting atomic heterogeneous operations via the EVM.


Flexible verification

Rather than committing to a single verification approach, Ritual is proof system agnostic. We recognize that different applications have different requirements for security, cost, and performance. Our modular computational integrity system lets developers choose and combine:

  • ZKML for maximum security and privacy
  • OPML for efficient verification of larger models where interaction is acceptable
  • PPML for cheap, non-interactive statistical verification through backdoors
  • TEE for hardware-based security

The goal is to match the right verification strategy to your specific needs, not force one approach for all use cases.


Node specialization

Traditional blockchains like Ethereum operate by coming to consensus over ordering and replicated execution of transactions.

Ritual introduces node specialization through architecture purpose-built for reducing redundant re-execution and enabling user preference, letting Ritual scale special-purpose compute.

Node specialization

Nodes can specialize in select types of compute including AI Inference, ZK Proving, or TEE Code Execution, rather than servicing all operations, earning outsized rewards for performance.


Long-running, stateful execution

With the extension of execution sidecars to be stateful, Ritual will adopt a minimally invasive approach to enabling asynchronous execution for long-running compute precompiles, all orchestrated via the blockchain.

This will enable radically new applications to be built that out-source long-running, computationally-intensive tasks to a robust network of underlying compute providers.

Execution Sidecars

Optimized underlying architecture for any type of heterogeneous compute, including AI inference, ZK proof verification, TEE code execution, and beyond.


Web2 adapters

Ritual acts as a drop-in AI Inference API replacement for common Web2 use-cases, offering competitively priced, redundantly orchestrated, and private and verifiable compute.

Infernet ↔︎ Chain

Ritual, with its underlying mesh of 8,000+ connected Infernet nodes, lets anyone, anywhere tap into transparent AI, with strong privacy and verifiability guardrails.


Ritual-as-a-service

ritual-and-world

Ritual ↔︎ World

Not only will end-users consume applications natively through Ritual, but other blockchains themselves symbiotically benefit by tapping into Ritual compute.


Robust agentic primitives

Ritual built Frenrug, the first on-chain AI agent.

Now, we are laying the fundamental groundwork for AI agents that live, interact, and transact entirely on-chain: