# Google Document AI Alternative

**URL:** <https://discuss.huggingface.co/t/google-document-ai-alternative/109200>\
**Category:** 🤗Transformers\
**Created:** [September 28, 2024, 6:29pm UTC](https://discuss.huggingface.co/t/google-document-ai-alternative/109200 "2024-09-28T18:29:14Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![fakebizprez](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/fakebizprez/32/27276_2.png) [@fakebizprez](https://discuss.huggingface.co/u/fakebizprez)\
**Post date:** [September 28, 2024, 6:29pm UTC](https://discuss.huggingface.co/t/google-document-ai-alternative/109200/1 "2024-09-28T18:29:14Z")

</div>

I previously developed a custom document processor using Google Document AI for my business. Recently, I’ve established on-premises infrastructure to host our applications. Now, I aim to recreate locally the functionality of our existing processor, which currently parses and extracts information from PDF documents with 100% accuracy.

**I’m seeking recommendations for models and components available on Hugging Face that could help achieve this goal.** While I’ve experimented with various Python libraries, most were primarily focused on OCR rather than comprehensive document processing. I would greatly appreciate any suggestions, feedback, or advice on this matter.

Thank you for your assistance.

---

<div class="post-metadata">

**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](https://discuss.huggingface.co/u/John6666)\
**Post date:** [September 29, 2024, 6:41am UTC](https://discuss.huggingface.co/t/google-document-ai-alternative/109200/2 "2024-09-29T06:41:42Z")

</div>

I don’t know if this fits your application, but this is the most common model I’ve seen lately for OCR.

> **[stepfun-ai/GOT-OCR2\_0 · Hugging Face](https://huggingface.co/stepfun-ai/GOT-OCR2_0)**
>
> We’re on a journey to advance and democratize artificial intelligence through open source and open science.

> **[Tonic's GOT OCR - a Hugging Face Space by Tonic](https://huggingface.co/spaces/Tonic/GOT-OCR)**
>
> GOT - OCR (from : UCAS, Beijing)

---

<div class="post-metadata">

**Author:** ![fakebizprez](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/fakebizprez/32/27276_2.png) [@fakebizprez](https://discuss.huggingface.co/u/fakebizprez)\
**Post date:** [October 6, 2024, 10:47am UTC](https://discuss.huggingface.co/t/google-document-ai-alternative/109200/3 "2024-10-06T10:47:10Z")

</div>

Thank you, @John6666, I appreciate the response - I know OCR isn’t too exciting.

I just looked into both of these models, and it appears that PDF isn’t supported. I reached out to the developers to confirm.

Thanks again for looking out!

---

<div class="post-metadata">

**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](https://discuss.huggingface.co/u/John6666)\
**Post date:** [October 6, 2024, 10:55am UTC](https://discuss.huggingface.co/t/google-document-ai-alternative/109200/4 "2024-10-06T10:55:34Z")

</div>

If you’re using PDF for input, it looks like there are models like this. If you use it for output, maybe it is something that is usually converted manually.

> **[PDF OCR - a Hugging Face Space by pszemraj](https://huggingface.co/spaces/pszemraj/pdf-ocr)**
>
> Discover amazing ML apps made by the community

[https://mindee.github.io/doctr/index.html](https://mindee.github.io/doctr/index.html)