---
title: "Article text from Image"
video_id: HuSXYnPZhb0
url: https://www.youtube.com/watch?v=HuSXYnPZhb0
workflow_page: https://workflows-of-ai.com/workflow/article-text-image
language: en
captions: generated
---

0:01 hey there Markus here next workflow is
0:04 the opposite of the last one um this is
0:07 not the workflow that makes sense but
0:11 it's um still one to show uh what you
0:16 can do in terms of image descriptions so
0:19 what we're going to do is we have a
0:21 header image but we don't actually have
0:23 an article so we're going to generate an
0:25 article from a header
0:27 image um and this is going to I use
0:30 something called xenex that can
0:32 basically explain images as always this
0:35 is druple um this is the open AI
0:40 interpolator module if you want to know
0:42 more about them uh there's soon going to
0:45 be links in the description where you
0:47 can see how you install it uh set it
0:50 up um and the idea behind the
0:54 interpolator module um Dro and the
0:58 interat module is open source everything
1:01 is free everything you see here uh you
1:04 can just set up yourself after uh you
1:08 finished watching this
1:10 video so I'm just going to start by
1:13 logging in
1:15 here so we have a content type in this
1:17 case there is an
1:19 article but yeah we only have a header
1:21 image so we need to create well first of
1:24 all we need to have a field in between
1:26 that's basically the image description
1:30 so this is where CX is going to take its
1:33 image description and put it in here
1:35 basically so we just need one field not
1:38 the multi field full
1:40 HTML they don't write HTML out so we
1:43 could have used uh plain text as well
1:45 but anyway we're we're going to use that
1:48 full HTML here CeX image description is
1:52 the rule we want base mode and we take
1:54 it from header image uh and put batch
1:57 mode on so we can see progress
2:00 and then we basically need to add the
2:04 article
2:06 text and this should be formatted long
2:09 because we actually want to create uh
2:12 HTML in this case so we're going to
2:14 enable AI interpolator here as well
2:16 we're going to use open AI text log so
2:19 that's taking text and generating text
2:21 right we're going to use the image
2:23 description as the uh context and then
2:27 we just write our prompt So based on on
2:30 the context write a 10
2:35 paragraph article about the uh
2:40 subject you may use HTML tags like
2:46 headers um strong and
2:50 M please set
2:53 up
2:55 headers uh to have some structure in the
2:59 article
3:01 context is then we copy from
3:05 here and of
3:07 course um wait first before I start
3:11 talking about that we're going to use
3:12 batch mode GPT 4 creates better uh or no
3:17 actually we don't care so much about
3:19 quality now so if you want better
3:21 quality use gp4 but I want a a quick
3:23 result now you are an uh uh left
3:30 top journalist no what do you call it a
3:34 journalist that writes about tech so
3:38 that's the role we want we have to set
3:40 the interpolator weight higher than the
3:42 last interpolator weight because we have
3:44 to take the image to image description
3:46 before we can write the article because
3:47 otherwise there's no text to base the
3:49 article on
3:50 right um then we just save this and go
3:53 to our content and resave it so we see
3:55 now we have a image description and an
3:57 article but they're empty
4:03 um
4:05 right so what I was going to say of
4:07 course this this is not a a a good
4:11 um example of an actual workflow because
4:14 what's going to happen now is that the
4:16 CX will not produce that much text and
4:20 of course uh um of course open AI can
4:24 you know make this text more expressive
4:28 but unless you tell it to keep to a
4:31 specific context and stay within the
4:34 facts of that context it will make up
4:36 text because I wrote that I wanted 10
4:39 paragraphs right uh and because of this
4:43 uh it might hallucinate so you should
4:45 never do this unless you wanted to make
4:47 up stuff of course so here you can see
4:52 um yeah so this was not super good
4:56 actually C next is not always great so
4:58 it thinks it's actually actually MacBook
5:00 Pro and as far as I can
5:04 see this should be the MacBook
5:08 Air I don't know Apple products that
5:11 well to be honest no maybe that is the
5:13 MacBook Pro I'm actually not sure anyway
5:18 so this is the description that I
5:21 created and this is the article maybe we
5:23 can look at it like this
5:26 instead so here's the article right with
5:30 headers as we asked for the quality is
5:32 not great because I used 3.5 in this
5:34 case but you see it can create an
5:36 article
5:37 from uh the image
5:41 basically um but yeah it's a lot of
5:44 hallucination and so on but there is of
5:47 course a use a much clearer use case and
5:51 that's of course using this to uh create
5:54 all tags and so on um but I wanted to
5:58 show sh basically so that's why I I took
6:01 this
6:03 workflow um yeah that's it thank you for
6:07 watching