GPTits

v0.4.10 New free and offline mode: Local AI Abort generation button Works best with Nvidia GPU; let me know if you encounter any issues with AMD. Koboldcpp included, .gguf models not included. Horde with more information and settings. Improvements in the interface. Bug fixes. v0.4.9 Character page loading faster. Performance improvement Bug fixes. v0.4.8 Improved menu *continue* button New artificial intelligence for Patrons (limited tokens) Automatic settings saving (except for the character) Dynamic Scenario for OpenAi (changes the scene, character image, and background image) Chromakey Integration with Stable Diffusion API (Tested only with AUTOMATIC1111: ) Use of .json to save characters Compatible with most of these 912 characters: Button to turn messages into images (requires Stable Diffusion) Button to regenerate the message/image Edit button List of image generation buttons for each character. Patreon log in improvements Button to log in manually with a key Horde AI, Generate an API Key at this link, default 0000000000 has low priority Automatic chromakey if the image has transparency, you can filter which models to use with horde Auto Save and load session, max tokens increased, MhytoMax formatting, Julia removed 10 example characters: 2B, Ashley, Azula, Elizabeth, Lara Croft, Makima, Mona Megistus, Raven, Tristana, Widowmaker Minor fixes v0.3b In the "OFFLINE AI" folder, there is a "GGML.exe" that works on port 5001 (based on LostRuins/koboldcpp). "Harlen", New example character in Pygmalion format, should also work with other models too. Words for substitution need to have a "$" in front (e.g., $NAME). You can delete error messages. Minor fixes. v0.3 Free mode and Patreon mode (with limitations). To use the 16k model (for very long conversations), you need to use your own API key and UI settings. There's no need to open the config.txt anymore. Anonymous chat sharing. OpenAI is more restrictive about generating sexual content. Please use AnonShare to create a model or lore independent of OpenAI. Three languages: English (EN), Brazilian Portuguese (PT-BR), and Spanish (ES). Display web content compatible with PNG, JPG, WEBM, MP4, and GIF formats. Additional settings for OpenAI, including temperature, max_tokens, top_p, frequency penalty, and presence penalty. Hovering over a menu option provides an explanation. Cut-off text at the end is removed. A "Characters" directory to store your characters. Character creation with multiple parameters, accompanied by examples. Random sexual injection for heavier fetishes. It chooses one from all that you add. If the token count approaches 4097, it will automatically reduce to avoid errors. Clear chat button. Right-clicking on a message copies it to the clipboard. Popup when the left chat is hidden. Removes text within parentheses from TTS reading. Ability to change the voice to ElevenLabs by name instead of ID. Character images and videos. Remove and copy message buttons. Customize the character with a .PNG image. Six example characters. PNG and PSD templates for character creation. Animated or static character options. Proper display of non-Latin languages. Some things that I forgot. v0.2 Added button description on the bottom right You can use LLAMA offline, without using the OpenAI service (v0.2b) GPT error messages with more context (v0.2b) LLAMA_MODEL setting removed to increase speed , to change the model edit the model name in /LAMA/ main.py (v0.2b) included empty default "models" folder inside LLAMA (v0.2b) Fixed bug that prevented the use of LLAMA in some cases (v0.2b) included a backup of config.txt (v0.2c) the LLAMA_URL can be used with a URL without a PORT v0.1 Released