Karl PRO

## Built for AI/ML consumers Most public energy datasets were assembled for traders, regulators, or journalists — not for ML pipelines. Schemas drift, primary keys are ambiguous, joins require domain folklore. OilPriceAPI is built the other way around. The maintainer has been working in AI/ML since the 1990s — through expert systems, classical ML, the deep-learning wave, and now LLM agents — and the platform is engineered specifically for those consumers: stable schemas, normalized join keys, geocoded coordinates, deduplication at ingest, and a single canonical operator name per company. The result is data you can `load_dataset()` and feed straight into a pipeline without a six-week cleanup phase.