# Getting Started

Let's briefly preview the steps to generate and verify a model:

1. Model Preparation: prepare model and exports it as an ONNX file.
2. Settings Generation: run ezkl.gen_settings() to generate initial settings for the proof. Can optionally use ezkl.calibrate_settings() to optimize these settings.
3. Model Compilation: compile model using ezkl.compile_model(), which prepares it for the setup phase.
4. Setup: run ezkl.setup(), which defines the rules for the proof and verification. This step produces a proving key, a verification key, and circuit settings.
5. Witness Generation: generate the witness data (input-output pair) using ezkl.gen_witness().
6. Proof Generation: create the cryptographic proof using ezkl.prove(), which takes the witness data, compiled model, proving key, and other necessary components.
7. [OPTIONAL] Verifier Creation: If the proof needs to be verified on-chain, the user can create an EVM verifier contract using ezkl.create_evm_verifier().
8. [OPTIONAL] Deployment: For on-chain verification, the user deploys the verifier contract using ezkl.deploy_evm().
9. Verification: The proof can be verified using one of three methods - on-chain using ezkl.verify_evm() with the deployed contract, off-chain via CLI using the ezkl verify command, or off-chain in Python using ezkl.verify()

Each of these steps and associated vocabulary are covered in detail in the following sections. Continue to the next section for installation.