| What you want to do | Project | What you get |
|---|---|---|
| Parse PDFs or document images | OXR | Markdown and JSON with layout coordinates |
| Classify a document or verify information on a page | Docision | A classification or verification result for each question, selected from the answers you provide |
| Answer questions about images, video, or audio | mJev | An answer to each question and scores for each option |
| Render PDFs, extract text, or split and merge files in Python | SoPDF | Page images, extracted text, or a processed PDF |
| Render Markdown with math or chemical structures | SoMarkDown | Rendered Markdown that you can display on a web page or in an app |
| Compare document parsers or versions | DocParsingBench | Text, formula, and table scores along with differences for each page |
Start here
Project overview
Find open-source models, developer tools, and benchmarks from the SoMark team by task.

