Roadmap

Live

Orgs, ducklings, API keys, AUTH_MODE. Files in R2. DuckDB in Containers (tested live). Optional WASM. HTTP ingest + compact. R2 Data Catalog on the bucket.

Next

1. Bind Pipelines → POST …/events
2. R2_SQL_TOKEN + one documented big scan
3. Catalog token + Iceberg tables end-to-end
4. Cleaner Container lifecycle
5. CI smoke on staging

Out of scope here

Warm pools that survive Container sleep, Spark-grade ETL inside a duckling, hosted SLA packaging. Batch into R2 (or use a warehouse built for that) and query from this stack.