We recently launched ColiVara - a state of the art RAG API that is based on vision LLMs with a state of art performance on both text-heavy documents and visually-complex pages.

It outperforms OCR-based pipelines by up to 30%+ on public benchmarks Benchmark Comparison Chart

We thought it might be of interest to you. If you're looking/exploring or know others who are we'd love for you to try it out.

Try it out: ColiVara.com

View the code: https://github.com/tjmlabs/ColiVara