Skip to main content
Runtime Plan

Datasets

Production data, curated

Build evaluation datasets from real production traces. Annotate, version, and use for benchmarking and fine-tuning.

From Production

Select traces from production to build datasets. Real-world data, not synthetic.

Annotation Workflows

Collaborate on labeling with your team. Built-in annotation tools and review flows.

Versioning

Every dataset is versioned. Track lineage and reproduce experiments exactly.

Full Capabilities

Build from production traces
Manual and bulk annotation
Custom label schemas
Dataset versioning
Export to CSV, JSONL, Parquet
Fine-tuning data formatting
Split into train/test/val
Lineage tracking

Use Cases

Build golden datasets for evaluation

Create fine-tuning datasets from production

Benchmark new models against real data

Track dataset quality over time

Unlock Datasets

Upgrade to Runtime starting at $149/month for the complete Waxell experience.