Lots of good vocabulary to learn:
tumescent. satiate. perineum. areolae. erectile.
Free NSFW uncensored AI chatbot
- Thread starter FAXz
- Start date
tumescent. satiate. perineum. areolae. erectile.
issue report: I tried using it with firefox browser on a android phone. The input text box would not accept input from the "gboard" keyboard "suggestions" bar - it only worked if every letter was typed out one at a time. It also failed with swype style input. there was no error message in either case, it just ignored the input.
Trying to use voice input failed as well but at least there was a "this app does not support voice input" error message, although i suspect that was due to the incognito mode.
Trying to use voice input failed as well but at least there was a "this app does not support voice input" error message, although i suspect that was due to the incognito mode.
It keeps popping an Error 500. Would it be feasible to run it on a local instance, like Ollamma or Oobabooga?
This specific one is a web app that OP is running with some shared LLM backend, 500 suggests something is borked with the app. Not sure what model he is using.
Lots of info on running your own local setup here: https://f95zone.to/threads/ai-uncensored-text-generation-via-oobabooga.162713/
that would be when the webserver is working to give you the html/JS frontend, but the api server is not responding.
- Jun 5, 2020
- 251
- 404
It's already down? Been 2 days (since about the same time as pipopipa123456) and it's still
Error fetching models.
Make sure your OpenAI API key is set in the bottom left of the sidebar.
If you completed this step, OpenAI may be experiencing issues.
Code: 500
Error fetching models.
Make sure your OpenAI API key is set in the bottom left of the sidebar.
If you completed this step, OpenAI may be experiencing issues.
Code: 500
- Jul 7, 2017
- 80
- 770
Faraday locally on your computer for chatting. Free, no filters. Faraday makes it easy to run open source LLMs (chatbots) on your computer. Once you have installed the program and AI models, no internet connection is required to use and interact with the LLM AI. Faraday.dev supports a wide range of LLaMA-based models, including WizardLM, GPT4-x-Alpaca, Vicuna, Koala, Open Assistant, PygmalionAI and others. You have the option to choose which models to download and use in the application.
You must be registered to see the links
It hasn't been working for month now and back then you didn't need a key.
I would assume someone pulled the plug because people used it either too much and it costs the owner money or people used it for lewd stuff that wasn't okay with whoever provided the AI (like underaged stuff).
But there is a Discord link, maybe there's more information in there.
Hey guys. sorry for late update. as you might to know, ai stuff costs money and its pretty hard to sustain such project. i tried to find some financial investments but failed.
Statistic (amount of visitors, amount of requests to ai) was pretty impressive for the project without any investments in ads. It looks promising, so i'd like to continue but (as i said earlier) its costs too much money for pet project...
I can revive it and introduce some kind of subscription (using crypto) to make financial situation a bit easier to me. What do you think about it, guys? Is it worth it?
Statistic (amount of visitors, amount of requests to ai) was pretty impressive for the project without any investments in ads. It looks promising, so i'd like to continue but (as i said earlier) its costs too much money for pet project...
I can revive it and introduce some kind of subscription (using crypto) to make financial situation a bit easier to me. What do you think about it, guys? Is it worth it?
There's now a lot of slick fancy sites around that are doing something similar. all of them seem to have various complicated hustles around payment and tokens and invitations etc etc. It feels like marketing and the sort of bullshit that EA plays with game skins/season pass etc. I don't like that at all.
If I had the power to make something perfect, i'd want something that was simpler visually but more featureful:
1. clear free/paid tiers and clear costs to use (5$ a month patrons get x00's of generations)
2. Really good tools for writing/editing/guiding the chat exactly the way you want. The UI you had before was okay but still was a bit clumsy to edit/revert/clear responses where the bot started going off track.
Using these sort of ai chat tools, I find I get part way through a chat/story and suddenly the bot would lose the plot and I would have to backtrack a few responses to get it back again.
It's clear the bots constantly need personality and scene hints/prods, but that stuff doesn't belong in the story flow itself.
So to combat this, I'm imagining something like a UI where you could strike out/cut off the response text with a simple mouse stroke, and then replace it or add a "commant hint" that is not part of the story, and continue the generation from where it left off.
In my mind, for the "best" possible interactive chat, the interface should be semi-automatically fixing up the input token stream to include:
a) the personality /background info,
b) then the "what's happened so far" for continuity,
c) then as much as possible of the "immediate previous words/actions",
d) then "invisible hints" that guide the next token generations but do not end up in the visible output stream
The way that Silly tavern works is somewhat like this, but it's overly complicated in the UI and stuff has weird non-obvious names about how it fits together. The holy grail is to have multi passes so the bots can do the "continuity compression" summarization themselves, but i'm not sure the system is there yet. maybe as token scopes grow it will no longer be a problem.
3. A curated list of base characters/scenarios. The sites that allow unrestricted user creation end up as stinking cesspits of "grimory is your gloomy bullied girlfriend #328" with random anime pics.
3a. once the user picks a character, it should be possible and easy to modify the starting personality/scenario and save that for their own use
4. text input assistance. when really getting into a chat session it's somewhat difficult to type with one hand (lol). not sure how to solve this
Well, there's my brain dump for the day.
BTW thanks for putting the server up for all of us, at your own personal expense. It was fun.
If I had the power to make something perfect, i'd want something that was simpler visually but more featureful:
1. clear free/paid tiers and clear costs to use (5$ a month patrons get x00's of generations)
2. Really good tools for writing/editing/guiding the chat exactly the way you want. The UI you had before was okay but still was a bit clumsy to edit/revert/clear responses where the bot started going off track.
Using these sort of ai chat tools, I find I get part way through a chat/story and suddenly the bot would lose the plot and I would have to backtrack a few responses to get it back again.
It's clear the bots constantly need personality and scene hints/prods, but that stuff doesn't belong in the story flow itself.
So to combat this, I'm imagining something like a UI where you could strike out/cut off the response text with a simple mouse stroke, and then replace it or add a "commant hint" that is not part of the story, and continue the generation from where it left off.
In my mind, for the "best" possible interactive chat, the interface should be semi-automatically fixing up the input token stream to include:
a) the personality /background info,
b) then the "what's happened so far" for continuity,
c) then as much as possible of the "immediate previous words/actions",
d) then "invisible hints" that guide the next token generations but do not end up in the visible output stream
The way that Silly tavern works is somewhat like this, but it's overly complicated in the UI and stuff has weird non-obvious names about how it fits together. The holy grail is to have multi passes so the bots can do the "continuity compression" summarization themselves, but i'm not sure the system is there yet. maybe as token scopes grow it will no longer be a problem.
3. A curated list of base characters/scenarios. The sites that allow unrestricted user creation end up as stinking cesspits of "grimory is your gloomy bullied girlfriend #328" with random anime pics.
3a. once the user picks a character, it should be possible and easy to modify the starting personality/scenario and save that for their own use
4. text input assistance. when really getting into a chat session it's somewhat difficult to type with one hand (lol). not sure how to solve this
Well, there's my brain dump for the day.
BTW thanks for putting the server up for all of us, at your own personal expense. It was fun.
Last edited: