โใใใใๅ ็ใใขใชในใฏๅ ็ใๅพ ใฃใฆใใพใใใโ
English | ็ฎไฝไธญๆ
Aris is a telegram chatbot which supports popular LLMs and customized character presets. In the future will add support for more language models.
Model Name | Support | Custom Prompt | Image Output |
---|---|---|---|
Gemini Pro | โ | โ | ๐ซ |
gpt-3.5-turbo | โ | โ | ๐ซ |
gpt-4 | โ | โ | ๐ซ |
New Bing | โ | ๐ซ | โ |
Google Bard | โ | ๐ซ | โ |
Claude | โ | ๐ซ | ๐ซ |
Note: Bard cannot generate picture by prompt, it can only send pictures fetched from Google search.
Aris supports importing add-on presets from template, which makes it pretty easy to contribute your own character preset to her, even for those who have no coding basis.
To add your own preset to Aris, follow the steps below:
- Create a new text file (.txt, .py, etc.) on your local machine.
- Copy texts from the template and paste it into the file you just created.
- Edit your prompts following the instructions in the template, just like how you create your own custom preset using the bot's command menu in Telegram. Refer to other currently available add-on presets if there's any confusion.
- Submit your preset by creating an issue, either attach the text file (recommended) or paste the content of your preset into the issue description.
Alternatively, if you have experience in git, you can also fork this repo and submit your preset by creating a pull request. Just put your preset.py
file under /presets
directory and make sure to the file name same as the id
of the preset.
-
Download source code and go to project root directory.
git clone https://github.com/HanaokaYuzu/TendouAris.git cd TendouAris
-
Rename
.env.sample
to.env
and fill the values. -
(Optional) To enable Bing model, export cookies from https://bing.com/chat and save it as
bing_cookies.json
under\srv
directory. Sample file is provided for format reference. -
(Optional) To enable Claude model, export cookies from https://claude.ai and save it as
claude_cookies.json
under\srv
directory. Sample file is provided for format reference. -
Build docker container.
docker-compose up -d --build docker image prune -f # Remove unused dangling images (optional)