GPT-4 Vision + Zapier + MindStudio (INSANE Automations)

preview_player
Показать описание
Let's build an AI workflow where you can upload handwritten notes, have GPT4 vision read and transcribe the notes, and then email those notes to anyone you want.

Join My Newsletter for Regular AI Updates 👇🏼

Need AI Consulting? ✅

Rent a GPU (MassedCompute) 🚀
USE CODE "MatthewBerman" for 50% discount

My Links 🔗

Media/Sponsorship Inquiries 📈
Рекомендации по теме
Комментарии
Автор

The AI fatigue is real... maybe i'll create an AI workflow for working through wanting to choke the next AI app interface I see.

avi
Автор

great video, thank you!

08:04 also, congrats on hiring your new video editor!!

WesRoth
Автор

I love that you're building useful things ❤

j.hanleysmith
Автор

You're getting pretty good at releasing videos on exactly what im trying to learn right as im trying to learn it lol

kylequinn
Автор

I would love to see a video comparing 4090 to M3 MAX or M2 Ultra for running LLMs locally

seenox_
Автор

Matthew “who is also the sponsor of this video” Berman

rawper
Автор

Thanks, Matthew for the video. I was getting an error in my output and I discovered that in the Run Function for the chat GPT4, I have one more variable which is the GPT-4 Vision instructions which does not show in your video. I added the instruction "Generate/extract all text from the image" and it extracted the notes successfully. Keep up with the great content.

PabloCarmonapc
Автор

Pretty good. You might want to have the termination block be a revise document instead of chat.

Franchise-infoCa
Автор

Never used zapier before but I thought it definitely can do this just with GPT-4 vision. I'm using n8n at the moment, it's pretty cool and free to use

legendarystuff
Автор

Great Video, Thanks buddy 👍👍

Love from India 🇮🇳

artsofpixel
Автор

Gotta be a way to manually do this offline using ai too i expect a part 2 to this story :)

JNET_Reloaded
Автор

Hallo Matthew, I am really enjoying your contributions. Imo you do a lot of very good work for the community. Thumbs up, Bro!

timschannel
Автор

How can this workflow be modified to watch a Google drive folder for new image pdf files and extract specific data from those files and save that data into an Excel sheet in another Google Drive folder, using GPT4 Vision, please?

Great_Muzik
Автор

Matthew, could you tell us your choice for legitimate work tasks and complex automations— or provide a couple and explain your reasoning. Thanks for the videos.

eIicit
Автор

Very cool! Reminds me a little bit of Node-Red / Zapier!

BTW, how much does it cost to use GPT4-Vision? For example, does asking it to transcribe an image count as 1 request towards the 40 requests/hour?

I have some tables I'd like to extract from some poor-quality scans.. the tools I've tried thusfar (paddle, ocrmypdf, Pytesseract, easyocr, keras_ocr, etc) seem to have trouble with it. Would love to see if this new world of AI can do it, but don't want to invest a lot of money before I know it works.

bennguyen
Автор

Please do more automations about how to be more productive

PromovareUTOPIAN
Автор

Requesting a video on chat-with-code setups, ideally using local a LLM. I tried "talk-codebase" and it gave me errors when trying to use the local LLM feature so that would be a great video for unique content if you can use your excellent installation skills to show how to get it working.

MattJonesYT
Автор

🎯 Key Takeaways for quick navigation:

00:00 📝 *Introduction to the AI-Powered Workflow*
- Introducing an AI workflow for transcribing handwritten notes and distributing them via email.
- The workflow involves Mind Studio, Zapier, and GPT-4 Vision.
- Mind Studio, sponsored by uaii, is a central tool in this process.
01:23 🖼️ *Setting Up User Input for Image Upload*
- Demonstrating how to set up user input for uploading an image of notes in Mind Studio.
- Creation of a 'user input' node to allow image upload.
- Customizing the input interface for better user guidance.
02:06 🔍 *Implementing GPT-4 Vision for Image Analysis*
- Using GPT-4 with Vision to interpret and transcribe the uploaded image.
- Requirement of an OpenAI API key for this function.
- Integration of GPT-4 Vision into the Mind Studio workflow.
03:40 🌐 *Integrating Zapier for Workflow Automation*
- Utilizing Zapier for automation and distribution of the transcribed notes.
- Creating a 'Zap' in Zapier to catch and process data from Mind Studio.
- Explanation of setting up Zapier webhooks for data transfer.
05:06 ✉️ *Configuring Email Distribution with Gmail*
- Setting up Gmail in Zapier for emailing the transcribed notes.
- Demonstrating the process of sending emails to a predefined list.
- Testing the full workflow from image upload to email distribution.
06:42 🚀 *Finalizing and Testing the Complete Automation*
- Final steps in setting up the automation, including user feedback.
- Renaming and publishing the complete workflow in Mind Studio.
- Testing the entire process, from image upload to receiving the email.

Made with HARPA AI

warezit
Автор

Where you can choose "from value", does this mean that this will be used for spoofing?

PatrickDodds
Автор

A bit easy method what we do is download the transcript of the meeting and then ask GPT to generate meeting minutes.

debakarr