canary
canary copied to clipboard
add batching in ask response stream
currently we are using model that is "fast enough", so not needed at the moment