Aleix Conchillo Flaqué bff5b3e562 LLMAssistantContextAggregator: always aggregate TTSTextFrame ...

This is the assistant aggregator and we should add everything that is being
spoekn. Because of that we should use TTSTextFrame because those are the frames
that are actually spoken. We should send the aggregation as soon as the bot
stops speaking.

So, we don't need to handle `LLMFullResponseStartFrame` and
`LLMFullResponseEndFrame` anymore.

2025-02-20 17:41:31 -08:00

26 KiB

Raw Blame History

View Raw