# MAKE AI FLOW OCR reading **URL:** \ **Category:** Beginner Questions\ **Tags:** mapping, ai\ **Created:** [February 5, 2026, 8:15am UTC](https://community.make.com/t/make-ai-flow-ocr-reading/102829 "2026-02-05T08:15:07Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![Marcin\_Kasierski](https://avatars.discourse-cdn.com/v4/letter/m/bc79bd/32.png) [@Marcin\_Kasierski](https://community.make.com/u/Marcin_Kasierski)\ **Post date:** [February 5, 2026, 8:15am UTC](https://community.make.com/t/make-ai-flow-ocr-reading/102829/1 "2026-02-05T08:15:07Z")
### 🎯 What is your goal? Want to read OCR file ### 🤔 What is the problem? Hi everyone, I have roughly the following flow: The source is a PDF file, I read the content from it, create diagrams (something like SketchWow), send them for approval on Slack, and then publish on Instagram. I have only one problem — actually, I’m stuck at the very beginning. I already have a ready OCR file on Google Drive, but I’m not really sure what to do next: which module should I add to read that OCR so that I can then reference the extracted text, for example in OpenAI, to generate a prompt. Because in the OpenAI modules themselves there is nowhere to map an OCR file from Google Drive — I couldn’t find such an option anywhere. As a result, the Compose String always returns a value with length 0, no matter what I use, and I’m not really sure how to handle or fix this. ### 📸 Screenshots: scenario setup, module configuration, errors ![OCR_flow.jpg](https://europe1.discourse-cdn.com/flex013/uploads/make/original/3X/2/b/2bac07a8c72ffd687d6f3dad9ff7b10684cdee1f.jpeg) ---
**Author:** ![Stoyan\_Vatov](https://avatars.discourse-cdn.com/v4/letter/s/5daacb/32.png) [@Stoyan\_Vatov](https://community.make.com/u/Stoyan_Vatov)\ **Post date:** [February 5, 2026, 8:18am UTC](https://community.make.com/t/make-ai-flow-ocr-reading/102829/3 "2026-02-05T08:18:13Z")
Hey there, You mean you have a pdf you want to do OCR on? There are plenty of apps to do that, just type pdf in the search. I personally use [pdf.co](http://pdf.co) for OCR. If you insist on using chatgpt, then you need to upload the file first. But I suggest using a dedicated app built for doing pdf OCR than a generic llm. ---
**Author:** ![Marcin\_Kasierski](https://avatars.discourse-cdn.com/v4/letter/m/bc79bd/32.png) [@Marcin\_Kasierski](https://community.make.com/u/Marcin_Kasierski)\ **Post date:** [February 5, 2026, 8:24am UTC](https://community.make.com/t/make-ai-flow-ocr-reading/102829/4 "2026-02-05T08:24:59Z")
No, Google Docs does the OCR automatically for me. The problem comes later — what should I do so that the OpenAI module, using that file (based on the OCR file), generates a specific prompt? Because everywhere in the OpenAI modules I don’t see any option to add or map that OCR file. It doesn’t necessarily have to be OpenAI — it could be something else — but I’d like it to generate a specific JSON based on the knowledge from the OCR file, which I can then export further to other tools or applications. ---
**Author:** ![Stoyan\_Vatov](https://avatars.discourse-cdn.com/v4/letter/s/5daacb/32.png) [@Stoyan\_Vatov](https://community.make.com/u/Stoyan_Vatov)\ **Post date:** [February 5, 2026, 8:25am UTC](https://community.make.com/t/make-ai-flow-ocr-reading/102829/5 "2026-02-05T08:25:57Z")
Yeah its a separate module that uploads the file first and returns a file id. Then you use the file id in the prompt to reference it. ---
**Author:** ![Make\_Bot](https://dub1.discourse-cdn.com/flex013/user_avatar/community.make.com/make_bot/32/14661_2.png) [@Make\_Bot](https://community.make.com/u/Make_Bot)\ **Post date:** [May 6, 2026, 7:33am UTC](https://community.make.com/t/make-ai-flow-ocr-reading/102829/6 "2026-05-06T07:33:16Z")