Alloovium is built on the shoulders of outstanding open-source software. This page lists the projects whose licenses ask for attribution, along with how we use them.
The acknowledgements below cover open-source components used in Alloovium's document processing services. Each entry links to the upstream project and names its license. This list is reviewed when we upgrade or replace pipeline components.
Alloovium uses MinerU for document parsing within our extraction pipeline. Portions of drawing and document layout analysis are powered by MinerU and the MinerU2.5-Pro model from OpenDataLab.
Project: https://github.com/opendatalab/MinerU
License: MinerU Open Source License (Apache 2.0-based)
Document conversion and layout understanding for text documents uses the Docling toolkit and the granite-docling models.
Project: https://github.com/docling-project/docling
License: MIT License (models: Apache 2.0)
High-throughput model serving for our self-hosted extraction models.
Project: https://github.com/vllm-project/vllm
License: Apache 2.0
Book a demo to get started