Deep dives on the parts of faucet-stream that are a genuine engineering contribution — each grounded in the real implementation, with every claim verified against the code before it ships.
Masking, quality, contracts, schema-drift policy, column lineage, and SLAs enforced inside the pipe — per page, before any write — with a mask-first ordering guarantee proven in CI.
Read the paper →How faucet-stream runs Arrow end-to-end when both ends support it — negotiated as a pure escape hatch, with default builds byte-identical to the row path.
Read the paper →Advancing the replication checkpoint only from a durable bookmark — never from decoded WAL or a keepalive — so no committed change is lost on crash. Postgres, MySQL, MongoDB, SQL Server.
Read the paper →The reproducible 1M-row reference move: 712k rows/s, ~96× Meltano, single binary — the method, the caveats, and why streaming Rust wins.
Coming soonNo papers match those filters.
Install the CLI, scaffold a config, and move real data — nothing external to stand up.
curl -LsSf https://github.com/faucet-hq/faucet-stream/releases/latest/download/faucet-cli-installer.sh | sh brew install faucet-hq/faucet-stream/faucet-cli