[Release] ChatGPT + Sex Toys = Ethumia 1.2: Webtease with AI

[TheGreyOne]

If you are using Brave browser the Ollama connection will not work (happy to receive input on how to fix this). Switched to Chrome - works so far.

[kinkylucy]

If you are using Brave browser the Ollama connection will not work (happy to receive input on how to fix this). Switched to Chrome - works so far.

I've had a good look at this, and Brave doesn't give useful reasons as to why they block it. Just giving "FAILED" with no further information. I've tried to debug it, but cannot find any flags that fix it, unfortunately.

[Bumbum]

Try making your generation sizes smaller 512x1024 and reducing the steps to like 15.

This reduced the image generation time to around 15 seconds which helps. But most scenarios create 2 or more pictres at once which again makes it very long to wait. If you type to fast and not all images are generated before you send your message to the LLM , the tease may get stuck or confused most of the time.

And I tried it with a FOC and restim which is also a problem on low power systems like mine. The signal generation was "shaky" and not stable at all. But I also tried the different patterns beforehand and find there is not much difference in the feelings compared to the background stim. As far as I have checked for the reason is that Ethumia is not controlling the volume of restim?
I hope to have some time to try stereostim if that's better.

Which system are you using and what are your image generation times?

[kinkylucy]

Try making your generation sizes smaller 512x1024 and reducing the steps to like 15.

This reduced the image generation time to around 15 seconds which helps. But most scenarios create 2 or more pictres at once which again makes it very long to wait. If you type to fast and not all images are generated before you send your message to the LLM , the tease may get stuck or confused most of the time.

And I tried it with a FOC and restim which is also a problem on low power systems like mine. The signal generation was "shaky" and not stable at all. But I also tried the different patterns beforehand and find there is not much difference in the feelings compared to the background stim. As far as I have checked for the reason is that Ethumia is not controlling the volume of restim?
I hope to have some time to try stereostim if that's better.

Which system are you using and what are your image generation times?

The restim integration is undergoing significant work to enhance its customizability and overall quality. It'll eventually be the flagship integration for estim. We're in alpha, so these integrations are quite early. This is the first site ever afaik to directly integrate with restim (not via intiface) so I wanted to prove it's possible first.

The issue with scenarios needs to be resolved yes, its an issue with how I wrote the scenario not inherent to them, but it'll be resolved in the next update globally anyway.

I'm on a 4090 and I get images within 15 seconds with 27 steps at 1592x1156.

[kinkylucy]

V1.0 Released
++ Simpler Setup
Free accounts have 200 LLM Generations per month, until you need to setup ollama or upgrade.
AI Image Gen can be disabled or ignored.

++ EOS & GuideMe Support
- Uses AI to turn any EOS and GuideME tease into a sex toy enabled experience.
- Upload EOS & GuideMe teases
- Load your own from your drive

++ Hypnosis Kink Added
- Add trigger words for on screen effects
- Give AI your triggers it can use in scenarios/chat (Supporter only)
- And so much more!

++ Chastity Kink Added
- Connect with Chaster or use the built in chastity lock
- AI can add time to your lock!

++ Restim Support Improved
Create Restim packs
Choose a pack

++ Coyote Support Improved
Should work now

++ Voice Control
Control Ethumia with your voice (Supporter Only)

And lots and lots of bugfixes and improvements.

[verynicekojak]

Hi,

I tried again and unfortunately failed again. Image generation is turned off, chat is turned off... so what is on? Can I do anything as guest? UI suggest that I can, but it is unclear how. I must admit that I didn't try very hard, but without clear indication how much work is required to start anything, it is difficult to motivate myself. Do I need access to other AIs? Do I need to pay for them? I am a programmer and started some local LLMs myself, but still it is unclear for me what is going on.

BR,
VNK

[SubZ]

alright so, this is a cool and expansive format. As a AI noob i really appreciated the setup wizard. I still struggled with the setup regardless. Ironically i used ollama to help me set up the properties it took me a mintue to realize that the variables needed to be separated ex "ollama_origins = *" is actually VARIABLE NAME(ollama_origins) = VARIABLE VALUE (*). asking ollama for help let me realize this. I'm really a scrub when it comes to utilizing the systems that run my computer but with the ai and the setup guide i was able to get the local ai host running on the website.

as far as the image generation goes i got stuck. I have a AMD GPU and had a difficult time getting the comfyui to run i was able to get the portable cpu version to run but it would return
"request with non matching host and origin localhost != ethumia.com, returning 403"
I could not get the GPU version to run even after following several guides. this is probably the sticking point for me as after several attempts and hours i couldn't get it to go. I'll continue to strive towards getting to work but it really shouldn't be this obtuse if you want more users on this website. This site seems geared towards user contribution, but without user integration i don't see any collaboration. this really needs to be streamlined more to get more users.

so while this is exciting and the setup guide eases some of the frustration its still crazy difficult for someone who isn't a power user to set up.

[SubZ]

also after the current update i get a "Failed to load scenario: The quota has been exceeded." after trying to load the non ai tease edge for gloria. not sure if im doing something wrong

[kinkylucy]

also after the current update i get a "Failed to load scenario: The quota has been exceeded." after trying to load the non ai tease edge for gloria. not sure if im doing something wrong

This has been fixed. We cache it in your browser so it loads faster in future, but some browsers don't like that. Now if it doesn't like that, it still loads.

[kinkylucy]

alright so, this is a cool and expansive format. As a AI noob i really appreciated the setup wizard. I still struggled with the setup regardless. Ironically i used ollama to help me set up the properties it took me a mintue to realize that the variables needed to be separated ex "ollama_origins = *" is actually VARIABLE NAME(ollama_origins) = VARIABLE VALUE (*). asking ollama for help let me realize this. I'm really a scrub when it comes to utilizing the systems that run my computer but with the ai and the setup guide i was able to get the local ai host running on the website.

as far as the image generation goes i got stuck. I have a AMD GPU and had a difficult time getting the comfyui to run i was able to get the portable cpu version to run but it would return
"request with non matching host and origin localhost != ethumia.com, returning 403"
I could not get the GPU version to run even after following several guides. this is probably the sticking point for me as after several attempts and hours i couldn't get it to go. I'll continue to strive towards getting to work but it really shouldn't be this obtuse if you want more users on this website. This site seems geared towards user contribution, but without user integration i don't see any collaboration. this really needs to be streamlined more to get more users.

so while this is exciting and the setup guide eases some of the frustration its still crazy difficult for someone who isn't a power user to set up.

Thanks for the feedback, I am still trying to see how I can continue to make it easier. I've added LLM generation for free for users to get started with.

[kinkylucy]

Hi,

I tried again and unfortunately failed again. Image generation is turned off, chat is turned off... so what is on? Can I do anything as guest? UI suggest that I can, but it is unclear how. I must admit that I didn't try very hard, but without clear indication how much work is required to start anything, it is difficult to motivate myself. Do I need access to other AIs? Do I need to pay for them? I am a programmer and started some local LLMs myself, but still it is unclear for me what is going on.

BR,
VNK

Edge for Gloria doesn't require AI, so you can play that without doing any setup at all.
EOS and GuideMe teases can be loaded in. If you login, they'll use the Ethumia Cloud LLM (if enabled in AI Settings) to generate stimulation to go along with the tease.

Thanks for your feedback though, I am working hard to make it simpler without reducing functionality so people can get started easier and can know what they can and cant run.

Teases now have a little marker saying what engine they run on (Ethumia, GuideMe, EOS) and whether they are LLM/AI free.

[Bumbum]

After your suggestions about image generation with my weak GPU, I have tried to use the teases a few times. But with no success. As LLM I use the open AI API. But after about 10 responses I get nothing more from the LLM. This was 3 or 4 times in a row now. (On different days and also with reboots of my computer inbetween)

As far as I understand this tease, I don't use any "expensive" ressources from your server (LLM runs on my paid openAI account, and image generation local with ComfUI).

Is there another limit I am missing or is there a bug?

[kinkylucy]

Is there another limit I am missing or is there a bug?

It's possible you're hitting a wall with OpenAI not liking the sexual content and refusing to generate. You should login and use Ethumia Cloud which is free LLM generation.

[Daragorn]

I really like the idea, and it is surely promising, but i am having a lot of troubles setting it up honestly.
The base default LLM, at least to me, seems quite slow, many times gives back an empty token with error, and is overall, quite repetitive and not too good. I tried to set up Ollama, but i can't get it to work at all, not even locally and i don't understand why.

The image generator, i managed to set it up to at least be working, but i am having distorced and bad images all the time, so i have no idea how to set it up to obtain decently looking images.

I would really like to try it more since the idea is really intriguing, especially for the toys controls....but i havent been able to set it up decently, sadly.

[Bumbum]

It's possible you're hitting a wall with OpenAI not liking the sexual content

Maybe that's the reason. I thought if I use the paid variant, this doesn't matter.

You should login and use Ethumia Cloud which is free LLM generation.

I tried that today, but the responses are very slow. It is so slow that is distracting from the plot.

To be on the safe side, I would like to mention that you should not misunderstand my answers. I think it's brilliant what you've created and I just want to give you feedback on my experiences. If something comes across as negative, it may be due to the translation. My mother tongue is not English.