dlt is an open-source Python library for loading data. Developers write pipelines in plain Python that move data from REST APIs, SQL databases and file stores such as S3, GCS, Azure and SFTP into warehouses, lakes and vector stores such as Snowflake, BigQuery, Databricks, Postgres, DuckDB, Iceberg and ClickHouse. dlt infers schemas, adapts when sources change and supports incremental loading. The library runs wherever Python runs. The company dltHub additionally offers a paid platform for managed deployment, scheduling, monitoring and scaling.
dlt is licensed under the Apache License 2.0; with the library, data stays in your own infrastructure. dltHub is headquartered in Berlin, with offices in San Francisco, Montreal and Amsterdam. On the dltHub platform each organisation picks a US or European data region when it is created, and the choice is permanent. Code, configuration, secrets and job outputs stay in the regional data plane, but the control plane (users, authorisation, orchestration metadata) is hosted in the US, and log output passes through the US even for EU workspaces.