# Finally trying out Agentic stuff and now the question is how do I get my text in the real world in to my computer… OCR, Vision, etc. — retained record

> Retained reddit record from reddit:LocalLLaMA.

Canonical: https://highsignal.app/data/records/bd593d94a77b2d2d

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:LocalLLaMA)  published 2026-09-23 02:27 ingested 2026-09-24 02:40

[original source ↗](https://www.reddit.com/r/LocalLLaMA/comments/1wntilc/finally_trying_out_agentic_stuff_and_now_the/)

## retained text



```
I’ve come to realize I’ve captured a lot of text with my phone. Sure it has auto OCR on phone but by the time it’s on my computer it’s gone. I rather automate the process. Debating between vision and OCR or both with agentic harness voting on which is likely better.

I currently plan to use pi.dev on a folder in my computer to convert each image into text then stitch it into one final document.

I’d also like some sort of image model that can take a crappy cell phone image and turn it into a nice document that looks scanned. It seems image edit models can work to some degree.

What are you all doing? What tools / models do you recommend?
```



document fetched 2026-09-24 00:45 · canonical https://reddit.com/r/LocalLLaMA/comments/1wntilc/finally_trying_out_agentic_stuff_and_now_the
