airbnb/chronon

Health v2 · maintenance only

Adoption not part of the score

1063 stars · 107 forks observed · 2026-10-06

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Chronon is an open-source data platform from Airbnb for computing, backfilling, and serving ML features. It handles batch and streaming feature computation, low-latency online serving, and point-in-time-correct training data generation with observability tooling.

Use cases

• serve low-latency features for online ML inference
• generate point-in-time correct training datasets
• backfill historical feature values over large time windows
• compute windowed aggregations from event streams and batch tables
• monitor online/offline feature consistency and data freshness

When to choose

• you need a feature store with guaranteed training/serving consistency
• your ML features require both batch and streaming computation
• you need scalable backfills for model training data
• you serve high-fanout feature sets at low latency

When to avoid

• you only need simple batch ETL without online serving
• your stack is not JVM/Spark-based
• you want a fully managed SaaS feature store rather than self-hosted infrastructure

Facets

framework · maturity active

etl streaming monitoring machine-learning caching machine-learning big-data artificial-intelligence jvm python cloud self-hosted feature-store feature-engineering online-serving backfills point-in-time-correctness scala ml-platform data-engineering docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="airbnb/chronon")