Evaluation API

The Evaluation API is the runtime interface for managed offchain risk policies. An application uses it to request a wallet risk evaluation and receive an onchain permit that authorizes a protected transaction.

Access to the Evaluation API is available after account provisioning. To get started, request access.

How it works

The Evaluation API follows a request-and-poll pattern:

1. Request an evaluation: Your application submits a transaction intent: the caller address, target contract, function selector, chain, and the permit parameters that will be extracted from the eventual onchain transaction.
2. Receive a permit ID: The API returns a deterministic permit ID for the evaluation. This ID is stable — submitting the same request again returns the same ID (idempotent).
3. Poll the evaluation status: The evaluation moves through statuses: evaluating, approving, and a terminal status (ready, rejected, or failed).
4. Submit the protected transaction — When the status becomes ready, the permit has been stored onchain by the CertifiedActionDONValidatorPolicy. Your application can now submit the protected transaction, which consumes the permit.

If the evaluation is rejected (for example, a screened address exceeded the configured risk threshold) or failed (for example, an internal error), no permit is created and the protected transaction will revert.

What you need

• A managed offchain risk policy configured and attached to your target contract.
• The permit parameters must correspond to supported extractor outputs and exactly match the values that will be extracted from the eventual onchain transaction.
• CRE Service Quotas apply to evaluations, including the HTTP trigger rate limit.

Related pages

• Off-Chain Policy Execution — conceptual overview of managed and custom offchain policies
• CertifiedActionDONValidatorPolicy — the onchain contract that stores and validates permits
• Reporting API — query permit delivery and evaluation outcomes