weberi / taskyto

A platform to develop LLM-based chatbots

Language and library for chatbot development

Usage

To run a specific chatbot there is a script (main.py) which is in charge of loading the chatbot and interpret the contents of the yaml files. The yaml files must be located in some folder. For example, examples/yaml/bike-shop. Then:

python src/main.py --chatbot examples/yaml/bike-shop/

Configuration

API keys can be set as environment variables or in a keys.properties files. The most important API key is OPENAI_API_KEY.

The keys.properties has the following form:

[keys]
SERPAPI_API_KEY=anapiforgooglesearch
OPENAI_API_KEY=theopenaikey

Other examples

There are some initial examples developed directly using the API. They are mostly incomplete and likely not working.

Bike shop

Requires OPENAI_API_KEY.

PYTHONPATH=src/ python3 examples/manual/bike-shop/bike_shop.py 

In Windows, use set PYTHONPATH=src before executing the command.

Healthy AI

A chatbot to answer questions about healthy habits.

To make this example work you need to manually add a book in PDF with the name data/body-building.pdf so that the system can index it to answer queries about this topic.

Requires SERPAPI_API_KEY and OPENAI_API_KEY.

PYTHONPATH=src/ python3 examples/manual/healthy-ai/healthy_ai.py