dltHub × LanceDB
LanceDB is an open-source multimodal vector database built for AI workloads — storing and querying vectors, images, audio, video, and structured data in a single columnar format based on the Lance format. It runs embedded (no separate server), scales to billions of rows, and integrates natively with PyArrow, pandas, and Ibis.
dltHub with LanceDB as a destination gives you a production-grade multimodal ELT pipeline: incremental ingestion from any REST API or data source, automatic schema normalization, deduplication, and state tracking — so your LanceDB tables stay fresh without reingesting the full corpus on every run.
Events