SuperwhisperSuperwhisper
Mar 21

Backup AI Model

When an AI model fails, it would be nice if there was a backup and you could set the amount of time before the backup is run. For example: if a response doesn't come back within two seconds for a short transcript, it would be ideal to switch to your backup model and run it instead. Sometimes things hang for 10 to 20 seconds, and I end up just having to go kill the application and start again. It should be able to recognize this and switch to a different model, or maybe even be able to run one in parallel so that you always get a response very quickly. I use a local AI model and it's extremely fast unless I am bogging it down by processing huge amounts of documents. In this case it would be ideal if after 1-2s it could simply switch to a cloud model.
PendingPending