"gpt-chatbot-loom" is a tool for creating various "figures" or "characters". These bots enhance the user experience through a local web interface, which is personalized by the name and function of the bots.
The project uses Python as the primary programming language and integrates the use of ChatGPT for bot creations and gradio for the web interface.
The project has just started. It's functional but still in its infancy.
The tool can be run using Python:
python -m pip install -r requirements.txt
python -m chatbot_loom
You can do it the easy way by running the following script:
./run.sh
The tool allows for the creation and usage of chatbots:
- Creating Chatbots - Bot creation is currently done manually by depositing a JSON file, which has to follow a certain schema. The path to this file is determined by the
BOTS_FILE
environment variable. If no file is found, a standard bot file is automatically created. The schema looks like this:
{
"type": "array",
"items": {
"type": "object",
"properties": {
"id": {"type": "string"},
"name": {"type": "string"},
"description": {"type": "string"},
"entrypoint": {"type": "string"},
},
"required": ["id", "name", "description", "entrypoint"],
},
}
- Using Chatbots - Chatbots are used via a web interface, where bots can be selected via tabs. To use the bots, an API key for OpenAI must be deposited via the
OPENAI_API_KEY
environment variable.
Want to contribute? Just create a pull request.