Making an Open Source Twitch VTuber with TTS and Language Model processing, and creating a universal Chat Platform for LLMs like Pygmalion, LLaMA and others.
- SillyLossy (Cohee) and his gracious sharing of his code for https://github.com/SillyLossy/TavernAI-extras and also his Misaka Chara Card and sprites.
- henk717 and his creation of the start runtime for Project Akiko.
- Trappu for both Mobius and the Akiko Chara Cards.
- AliCat for Harry Potter Characard.
- AVAKSon for Gilgamesh and his sprites.
- Suggestions Form: https://forms.gle/pahKwy73D6DzCaJV8
- Trello Board (Check here for what's being worked on): https://trello.com/b/muwhUyM9/main-app
- Discord Server: https://discord.gg/Pdhd7dEqHp
- Clone the repo to your local machine.
- Click the 'setup.bat' file and wait for the batch file to complete setting up your enviornment. After you run this once, just use Start.bat for startup.
- Navigate to the link displayed on the Node CMD. usually it is 'http://localhost:5173/'.
- Go to the 'Settings' tab and configure your Textgen Endpoint.
- Start chatting!
- Clone the repo to your local machine.
- Go into the 'setup.bat' (and or 'setup.sh' on Linux). For windows users change this:
start cmd /k "cd backend && pip install -r requirements.txt && python server.py"
to this
start cmd /k "cd backend && pip install -r full_requirements.txt && python server.py"
For Linux users change this:
cd backend && pip install -r requirements.txt && python server.py&
to this
cd backend && pip install -r full_requirements.txt && python server.py&
After running it should boot you into Akiko.
- Adding in modules.
Name | Description | Required Modules | Screenshot |
---|---|---|---|
Image Captioning | Send a cute picture to your bot! Picture select option will appear next to the text input box. |
caption |
|
Character Expressions | See your character reacting to your messages! You need to provide your own character images! |
classify |
Name | Description | Included in default requirements.txt |
---|---|---|
caption |
Image captioning | ❌ No |
classify |
Text sentiment classification | ❌ No |
Flag | Description |
---|---|
--enable-modules |
Required option. Provide a list of enabled modules. Expects a comma-separated list of module names. See Modules Example: --enable-modules=caption,sd |
--port |
Specify the port on which the application is hosted. Default: 5100 |
--listen |
Host the app on the local network |
--share |
Share the app on CloudFlare tunnel |
--cpu |
Run the models on the CPU instead of CUDA |
--summarization-model |
Load a custom summarization model. Expects a HuggingFace model ID. Default: Qiliang/bart-large-cnn-samsum-ChatGPT_v3 |
--classification-model |
Load a custom sentiment classification model. Expects a HuggingFace model ID. Default (6 emotions): bhadresh-savani/distilbert-base-uncased-emotion Other solid option is (28 emotions): joeddav/distilbert-base-uncased-go-emotions-student |
--captioning-model |
Load a custom captioning model. Expects a HuggingFace model ID. Default: Salesforce/blip-image-captioning-large |
--keyphrase-model |
Load a custom key phrase extraction model. Expects a HuggingFace model ID. Default: ml6team/keyphrase-extraction-distilbert-inspec |
Add any of those arguments to the start.bat
or start.sh
like this:
python server.py --enable-modules=caption,classify --listen
To make your site public, use this in the same file:
npx vite --host
- Horde Support. (Done)
- Fully Color Custom UI. (Done)
- Good Character Creation Page. (Done)
- Connect to Kobold, Ooba, horde, or use Akiko's own backend. With Pygmalion and LLaMA support. (Mostly Done)
- Emotion Detection. (Done)
- Long-term and Short Term memory summarization.
- Image captioning. (Done)
- Stable Diffusion (probably).
- Character Creation
- Character Card Import/Export
- Guide Page full of info for creating characters in every available style, AliChat, W++, and more.
- Character creation and character management with exports to Character Cards (TavernCardPNG). (Done)
- Chat with your favorite Character through a polished UI. (Done)
- Chat with multiple characters in character groups. (Done)
- Chat regeneration (swipes), Editable Messages, Impersonation, Branching Chats. (75% Done)
- Emotion detection with corresponding emotional display for set-characters (configuration required). **(Done)
- Send images to your character and have them recognize what the image is. (Done)
- Live2D models with emotion display. (50% Done)
- Text-to-speech with AzureTTS or (maybe) other alternatives. (this will also allow you to see lips moving on Live2D models.)
- Configure whether or not you want your bot to know what the current time of day is.
- Configure 'time since' to show your bot how long it has been since you last messaged them.
- Configure your relationship to the bot (Friend, Lover, Wife/Husband, Sibling, Custom Value, etc).
- Configure 'consent' for certain actions for more 'realism'.
This feature and it's components will be optional.
- User sign-on/sign-up.
- Allow authorized users to access application settings.
- Have a user profile custom to you and your other users.
- Be able to invite users to chat with you and your AI with user recognition.
- See other user's stats (how many messages they've sent, what their favorite Character is, etc.)
- See a Character's creator.
- Character sharing (or not).
- Create your own AI VTuber like Neurosama!
- See and manage Twitch Stats and Chats from inside Project Akiko UI.
- Manage AI output so it is Twitch filtered (no more holocaust jokes!).
- Chat with your bot through discord.
- Send images to the bot.
- Ask for images from the bot (spicy or not).
- Custom Chat Settings are also available on the discord bot.