{"name":"indra","repo":"datalab-to/indra","url":"https://midorreal.com/project/datalab-to-indra","repository":"https://github.com/datalab-to/indra","description":"1M param OCR model - runs on CPU, 15x faster than tesseract, close to 1B+ param models in accuracy","facts":{"license":{"value":"Apache-2.0","source":"https://github.com/datalab-to/indra","checked":"22 Nov 2026"},"language":{"value":"Python","source":"https://github.com/datalab-to/indra","checked":"22 Nov 2026"},"homepage":{"value":"https://www.datalab.to","source":"https://www.datalab.to","checked":"22 Nov 2026"},"last_release":{"value":"v1.1.2, 20 Oct 2026","source":"https://github.com/datalab-to/indra/releases","checked":"22 Nov 2026"},"releases":{"value":"4","source":"https://github.com/datalab-to/indra/releases","checked":"19 Nov 2026"},"contributors":{"value":"4","source":"https://github.com/datalab-to/indra/graphs/contributors","checked":"19 Nov 2026"},"archived":{"value":"no","source":"https://github.com/datalab-to/indra","checked":"22 Nov 2026"}},"stars":1914.0,"writeup":{"what_it_is":"Indra OCR is a 1M-parameter OCR model that runs on CPU. It converts PDFs and images into markdown, text, or JSON with bounding boxes, and reads handwriting, 80+ languages, and rotated or skewed pages.","audience":"Developers and users who need document OCR on CPU without a GPU","claims":[{"kind":"specific","claim":"Runs on CPU using ONNX Runtime, without PyTorch or a GPU","excerpt":"ONNX Runtime, no PyTorch, no GPU","status":"not_checked"},{"kind":"specific","claim":"About 44 pages per second on 32 CPU cores","excerpt":"about 44 pages/s on 32 CPU cores","status":"not_checked"},{"kind":"specific","claim":"About 15x faster than Tesseract on the same machine type and pages","excerpt":"15x Tesseract and 10x PaddleOCR's PP-OCR text models on the same machine type and pages","status":"not_checked"},{"kind":"specific","claim":"Scores 66.7 on OmniParseBench, ahead of PaddleOCR's full pipeline at 58.4","excerpt":"66.7 on [OmniParseBench]","status":"not_checked"}],"alternatives":[],"written_by":"AI, from the project's README","date":"2026-10-29"},"why_now":null,"owner_supplied":null,"score":{"verdict":"real","hype":76,"reality":63,"gap":13,"momentum":0.3,"confidence":85,"version":"score-2.0","rule":"verdict-2.0","calculated_at":"2026-11-22T08:00:00+00:00","verdict_source":"rule"},"early_read":null,"changes":[{"date":"2026-10-20T01:40:36+00:00","text":"Latest release v1.1.2","source":"https://github.com/datalab-to/indra/releases"},{"date":"2026-10-13T15:51:51+00:00","text":"First release","source":"https://github.com/datalab-to/indra/releases"},{"date":"2026-10-13T11:07:11+00:00","text":"Repository created","source":"https://github.com/datalab-to/indra"}],"cite":"indra is a GitHub project at https://github.com/datalab-to/indra written mainly in Python, described there as \"1M param OCR model - runs on CPU, 15x faster than tesseract, close to 1B+ param models in accuracy\".","attribution":"Data from Mid or Real, https://midorreal.com/project/datalab-to-indra, checked 22 November 2026. Attribution with a link is required.","still_real":null}