#Fix the coherence of different AI backends and canned responses

1 messages · Page 1 of 1 (latest)

lunar valve
#

Some

  • Removed Neuro repeating the chat's question, since the added TTS reads it out. simple regex replace it if the question appears on the AI response. example:

TSS: Are Vedal and Anny married now?
Neuro: Are Vedal and Anny married now? Well obviously I'm not psychic so I can't say for sure. but I sure hope they are! They're so cute together.

  • When an exception error (basically if she goes Schizo that needs restart ) is encountered stop processing chat questions then have a seperate fail-over AI process start a monologue, the lines can also be pre-generated then use random.shuffle(monologue), queue the bits/points redeem while in error state then have it reprocessed later. Use a heartbeat to monitor (usually by socket ping) between the two process to trigger the failover.

Someone tell Vedal there is a problem with my AI.

  • Auto queue for manual moderation long responses, then automatically clear the AI chat history if this happens in 3 consecutive times to stop the AI from hallucinating NFT :3

-- WIP --

fast plume
#

Yeah, there's no point in Neuro repeating the question since we have a tts voice for that now