Overview

Package overview and exports.

dash_evals - Evaluation framework for Dart and Flutter AI assistants.

This package provides tools for running evaluations using Inspect AI to measure model performance on Dart/Flutter tasks.

Configuration is resolved by the Dart CLI (devals) and emitted as JSONL datasets + a run manifest. The Python package reads the manifest and calls eval_set() directly.

Main entry point:

run-evals –manifest <path-to-manifest>