R-Think

Artifacts

Artifacts are the persistent records produced by reasoning execution. They include traces, evidence, metrics, and validation results.

Artifact Types

R-Think provides several built-in artifact types:

  • Trace - A step-by-step record of the reasoning execution.
  • Evidence - The data and metadata that support each decision.
  • Metrics - Performance and quality metrics for each step.
  • Validation - Records of all validation checks performed.

Capturing Artifacts

Artifacts are captured automatically during protocol execution. You can configure which artifacts to capture using the artifact configuration.

artifacts.ts typescript
import { Runtime, Artifact } from '@r-think/core';

const runtime = Runtime.create({
  protocol,
  artifacts: [
    Artifact.trace({ includeInput: true, includeOutput: true }),
    Artifact.evidence({ store: 's3://evidence-bucket' }),
    Artifact.metrics({ includeTiming: true })
  ]
});

const result = await runtime.execute(input);

// Access artifacts
console.log(result.trace.steps);     // Step-by-step execution record
console.log(result.evidence);        // Evidence records
console.log(result.metrics);         // Timing and quality metrics

Artifact Storage

Artifacts can be stored in memory, written to files, or persisted to external storage. Use artifact adapters to customize storage behavior.