VioneT

VioneT, 1 day ago

I was able to fetch it with the super fetch here: perchance.org/4cozwe59pi

I also didn’t have any CORS problems.

VioneT, 1 day ago

I’m also on Edge and no CORS issue … weeeeiirdd. I’ve also tried it in a profile without extensions (to see if any of my extensions messes with it) but it also goes through no problem. Same with Google Chrome, Firefox, Opera GX, and Opera Normal.

VioneT, 1 day ago

Are you referring to adding an Avatar to the character?

You can copy the data:image/jpeg;base64,/… of the image you generated outside of the ai-character-chat and paste it on the Avatar URL box on the character settings to add it.
https://lemmy.world/pictrs/image/3b103e6f-4bd8-42cf-884f-1045e9240a67.png

You can also upload the images in image hosting sites like Imgur, Catbox, etc,. and link the resulting image url (the one with the file extension on the end of the URL like imgur.com/imagename.extension) or upload the image in Perchance Upload and copy the link generated to the input box.

You can also change the ‘prompts’ that are used in the character chat image generation by changing these settings:
https://lemmy.world/pictrs/image/b93cbe1d-ea54-4981-ade1-69b35f74f5bb.png

VioneT, 1 day ago

Just to note, the AI cannot parse any image to reply with, so if you are looking for it to reply based on the image you sent, it wouldn’t be able to.

You can add images by adding a <img> element like so:

<span style="color:#323232;"><img src="image.url">
</span>

Where the image URL can be a data object or regular URL.

Also note that if you haven’t set it as a hidden from AI, the long image URL might be added to the AI’s memories.

VioneT, 1 day ago

Here is a video demo of it

If you just want to add images, that is the way. You don’t really alter the code, but you just sent the image as a message to the thread.

VioneT, 1 day ago

Just remember that if you have autoreply enabled. It would reply based on what it gathered from the text you have inputted and not the image that is displayed. From the video, I added an URL and it assumed that it was for a video game lmao.

VioneT, 1 day ago (edited 1 day ago)

EDIT: I’ve referenced you into this post.

I’m brand new to this forum so i’m not sure what its main purpose is yet, is it for people who are just users of Perchance, or is it a community of people who are helping to develop and improve it, or both?

This is a ‘technical/help’ related forum for the site i.e. help with the Perchance syntax and building generators. For more casual posts see Casual Perchance.

I wish there was a model that was trained on nothing but VTT tokens that could handle what I’m trying to do.

Unfortunately, Perchance doesn’t have a ‘model’ training option to get the AI images you want. But you can create your own ‘style’ by creating your own prompt (groups of words to describe the image you want to create).

I would suggest asking for help in creating the style on Casual Perchance or at the Perchance Discord, this forum is more for building generators and not prompting.

Also note, Perchance isn’t an AI-focused site, it has the AI only as an additional tool.

VioneT, 6 days ago

There is no Image to Image in Perchance.

VioneT, 8 days ago

Since you have both items in the region item. Currently you have it like:

<span style="color:#323232;">The [r = regions.selectOne, r.name] region contains [r.state]
</span>

But, since you have stored the item in r you can just swap them like so:

<span style="color:#323232;">[r = regions.selectOne, r.state] is in the [r.name] region
</span>

Also, a tip to get the name of the list (to not need the name property in the item) you can do:

<span style="color:#323232;">[r = regions.selectOne, r.state] is in the [r.getName] region
</span>

See Perchance Snippets - Traversing Hierarchical Lists.

VioneT, 8 days ago

You can get a consumable list of the states with r.state.consumableList, then you can have it like:

<span style="color:#323232;">[r = regions.selectOne, rscl = r.state.consumableList, rscl.selectOne] is in the [r.name] region, as well as [rscl.selectMany(3).join(', ')].
</span>

VioneT, 7 days ago

First, you need to create a new variable, then just like the first one, you only need to change the regions list to the rcl consumable list like so:

<span style="color:#323232;">[r = rcl.selectOne, r.state] is in the [r.name] region <br>
</span><span style="color:#323232;">[r = rcl.selectOne, r.state] is in the [r.name] region <br>
</span><span style="color:#323232;">[r = rcl.selectOne, r.state] is in the [r.name] region <br>
</span><span style="color:#323232;">[r = rcl.selectOne, r.state] is in the [r.name] region <br>
</span>

You can use other variables other than r.

VioneT, 10 days ago

@perchance - pinging dev

VioneT, 9 days ago (edited 9 days ago)

Previously it worked fine, but now it doesn’t. Here’s the code for the ElevenLabs API that I wrote. Here is a HAR file of the requests.

VioneT, 9 days ago (edited 9 days ago)

It seems to work now, though there are some inconsistencies with the chunk text arrangement, which causes the text in the stream to be quite jumbled. I’m looking into it now, I’ll update if it is still inconsistent with the order.

EDIT: It is inconsistent with the order of the chunks. Maybe there is a way to parse it in order? Currently I’m having it pushed into an array, then sort that array by the index, then join the sorted array to be a string before pushing it to the t2s, though it is still inconsistent and sometimes the streaming finishes and the text to be spoken is not yet queued up.

Here are the code hacks to re-sort the chunks to order lmao.

VioneT, 7 days ago

You have. On trying your code. it gave me the OUT OF ORDER CHUNKS!:

https://lemmy.world/pictrs/image/9f4522db-67cd-4717-868b-a73c3e1a8785.png
https://lemmy.world/pictrs/image/e352ef84-23d7-4a2d-a6d2-86e0ce887aab.png
Here’s the character I used with the custom code to check the out of order chunks.: Link to Character

VioneT, 7 days ago

Niiceeeeeeeeeee it is good now. Thanks again!

VioneT, 10 days ago

I’ve updated this neat little tool to allow copy and pasting from the generator items in the text area and it should neatly format it and show the generators:

perchance.org/generator-details-checker

VioneT, 11 days ago

Here is my current attempt in creating a bot (or simply applying the ai-text-plugin) on the comments-plugin: bot testing 01

VioneT, 7 days ago

See this silly video

Since you have a ‘mirror’ on the bottom one, the com.inputText is being applied on the last instance of the comments. Which is why it isn’t being mirrored on the top one.

VioneT, 7 days ago

You can check if the current user has submitted with comment.byCurrentUser

VioneT, 1 day ago

Also the message bubble buttons seems messed up 😅 after the addition of the reply button.

VioneT, 1 day ago

https://lemmy.world/pictrs/image/3b2be500-5dce-4c89-a414-e7bb32aab70f.png

From /generators page

https://lemmy.world/pictrs/image/2fbb563e-1a21-48df-9d38-12d45bdf6c29.png

From /hub page with Admin enabled.

VioneT, 1 day ago (edited 1 day ago)

Also there seems to be the ‘delete’ comment on replies (not sure if intended but clicking on it works to delete the reply and the mentioned comment):
https://lemmy.world/pictrs/image/0c8066d3-ff65-4525-8a5d-3db85da629cc.png

VioneT, 12 days ago

The max that the generator list request gives are up to 7 days only unfortunately.

You can just add keywords to your $meta description and the search should match it.