Currently a WIP, best to ask me for how to run it locally, but here's the general idea:
- Create an Azure OpenAI resource with a
gpt-35-turbo
model and atext-embedding-ada-002
model - Create an Azure AI Speech resource (optional - disable by setting
Speech:Enabled
tofalse
inappsettings.json
) - Create an
appsettings.Development.json
file in the AppHost project and populate the details from Azure Resources as per the empty ones in theappsettings.json
- Launch the app, navigate to the frontend -> management screen
- Wait for podcasts to load, choose the one(s) to index and click the Index Selected Items button
- Wait for a bunch of time (no UI progress, but you can click the "Retrieve Status" button to see the indexing status)
- Once indexing is done, go to the frontend
/
index, enter a prompt and see what happens