Datavolo

Datavolo is a dataflow platform for building multimodal data pipelines that feed generative AI systems. It is built on Apache NiFi and aimed at the unstructured data, such as documents, files and other non tabular content, that large language model applications depend on. Pipelines replace single purpose, point to point code with reusable flows that can be reconfigured from any source to any destination. The product is aimed at data and AI engineers. It ships with more than 300 connectors and processors, and additional processors can be written natively in Python or Java. Pipelines can be built in a no code visual interface or described in natural language. Lineage is built into every pipeline, so the path of the data can be inspected. The vendor positions it as a way to swap LLMs, chunking and parsing strategies or vector databases without rebuilding existing flows. The cloud offering adds extensions for retrieval augmented generation, document intelligence that parses unstructured documents, PII detection and sanitation, integrations with leading vector stores and AI systems, and Kubernetes orchestration operators. Datavolo Foundations provides support for Apache NiFi from the vendor's team, with Starter and Enterprise support levels that differ in nodes, support contacts and support hours. Enterprise and Cloud Enterprise plans are quoted through the sales team. The site announces that Snowflake has agreed to acquire Datavolo, so the product now sits inside that company's ecosystem.