Ritual enables expressive, heterogeneous compute via the EVM precompile contract interface. In contrast to Ethereum and other EVM-compatible blockchains today, that implement precompiles directly in their execution client codebases, Ritual’s compute precompiles are instead implemented as disjoint sidecars that run and scale parallel to an execution client. With this modular design, anyone that can architect a containerized sidecar can automatically have their computation be enshrined on-chain. You can learn more about this architectural decision here.

​
Supported sidecars

Ritual currently supports a range of EVM++ sidecars designed for specific use cases, including AI Inference, ZK Proving & Verification, TEE Code Execution, Agents, and Chain Abstraction.

Classical ML Inference

Leverage canonical tree-based and regression-based models for efficient and scalable ML computation.

LLM Inference

Leverage state-of-the-art LLM models for complex language tasks.

ZK Proving & Verification

Generate ZK proofs for heterogeneous computation, ensuring verifiable and tamper-proof computation.

TEE Execution

Perform secure computation in a Trusted Execution Environment (TEE), enabling privacy-preserving AI model inference, autonomous & self-evolving agents, external API calls, and more.

Chain Abstraction

Natively read (and in the future, write) state to and from other blockchain networks all from Ritual Chain.