CLI Reference
benchmarks
Benchmark management and accuracy measurement
$
limai benchmarks <subcommand>Run accuracy benchmarks, list benchmark files, and inspect results and metrics.
Options
--jsonOutput raw JSON--verboseEnable debug loggingSubcommands
limai benchmarks run <extractionSchemaId>
Start a new benchmark run.
$
limai benchmarks run <extractionSchemaId>Run a benchmark.
Options
--name <name>Benchmark name (required)--description <desc>Benchmark description--file-ids <ids>Comma-separated file IDs to include--config-file <file>JSON file with benchmark configurationlimai benchmarks files <extractionSchemaId>
List files available for benchmarking.
limai benchmarks files <extractionSchemaId>limai benchmarks results <extractionSchemaId>
Get detailed results for a benchmark execution.
$
limai benchmarks results <extractionSchemaId>Get benchmark results.
Options
--execution-id <id>Execution ID (required)limai benchmarks status <executionId>
Get the status of a benchmark execution.
limai benchmarks status <executionId>limai benchmarks metrics <executionId>
Get accuracy metrics for a benchmark execution.
limai benchmarks metrics <executionId>Examples
Run a benchmark on specific files
limai benchmarks run es_abc123 --name "v2 accuracy test" --file-ids file_1,file_2,file_3Run a benchmark with a config file
limai benchmarks run es_abc123 --name "full suite" --config-file benchmark-config.jsonList available benchmark files
limai benchmarks files es_abc123 --jsonCheck benchmark progress
limai benchmarks status exec_xyz789View accuracy metrics
limai benchmarks metrics exec_xyz789 --jsonGet detailed results
limai benchmarks results es_abc123 --execution-id exec_xyz789 --json