# clovaai/donut

Official Implementation of OCR-free Document Understanding Transformer (Donut) and Synthetic Document Generator (SynthDoG), ECCV 2022

Repository: https://github.com/clovaai/donut
Canonical: https://ross.abutalabs.com/products/donut
Homepage: https://arxiv.org/abs/2111.15664
Language: Python
License: MIT
License Family: permissive
Topics: document-ai, eccv-2022, multimodal-pre-trained-model, ocr, nlp, computer-vision
Last push: 2024-07-11T15:33:26+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1506, "days_push": 783, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 6919, forks 565 (observed 2026-08-28T04:09:51.405412+00:00)

## What it is
Donut is an OCR-free end-to-end Transformer model for visual document understanding tasks such as document classification and information extraction. The repository also includes SynthDoG, a synthetic document image generator for flexible multilingual pre-training.

## Use cases
- extract structured data from invoices without OCR
- parse receipts into JSON
- classify document images
- pretrain a document understanding model on synthetic documents
- build a document parsing pipeline in Python

## When to choose
- you need document information extraction without depending on external OCR engines
- you want to avoid OCR error propagation and reduce OCR compute costs
- you need multilingual or domain-flexible document pre-training via synthetic data

## When to avoid
- you need general-purpose text OCR for plain text extraction
- you lack GPU resources for training or inference
- you need actively developed features or frequent updates

## Facets
- artifact type: library
- maturity: maintenance
- function: ocr, machine-learning, deep-learning, nlp, computer-vision, data-generation
- domain: computer-vision, machine-learning, pdf
- platform: python, cross-platform
- tags: document-understanding, ocr-free, document-parsing, synthetic-data, transformer, eccv-2022, information-extraction, natural-language-processing, gpu

## Member repositories
- clovaai/donut (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:51.405412+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:41:26.021818+00:00, confidence not recorded.
  - readme: https://github.com/clovaai/donut (fetched 2026-08-28T04:09:51.405412+00:00, sha d537d720b719)
  - homepage: https://arxiv.org/abs/2111.15664 (fetched 2026-08-29T08:37:19.416310+00:00, sha bd2fe6bcc1a9)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T08:37:19.425885+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T08:37:19.430073+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T08:37:19.431919+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T08:37:19.428349+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.