Getting started¶
The easiest way to get started with GeniA is using Streamlit web app. Make sure you have python3 & pip3 installed, then run:
Then you can run the streamlit web app by:
You can also play with GeniA in a Terminal using GeniA 'local' mode.
Both quick start options provide a sneak peek, but GeniA is crafted for team collaboration and works best in Slack.
For simplicity, we recommend running it locally using Docker. If you want to run docker locally, please refer to the Installation section.
Open AI Azure deployment¶
When using Azure OpenAI, add those to your environment variables:
OPENAI_API_DEPLOYMENT=
OPENAI_API_TYPE="azure"
OPENAI_API_BASE=https://<your-endpoint.openai.azure.com/
OPENAI_API_VERSION="2023-07-01-preview"
Additional notes¶
Cost implications¶
GeniA uses OpenAI, be mindful of cost implications and ensure you set usage limits. You can configure both soft and hard limits at the following URL: https://platform.openai.com/account/billing/limits.
Model version¶
By default, GeniA is set to use gpt-3.5-turbo-0613. We acknowledge that gpt-4-0613 often delivers superior results, but have found the 3.5 version to be a more cost-effective choice.