* send one request for assign all clients
* tweak narrate-after-dialogue prompt
* elevenlabs default to turbo model and make model id configurable
* improve add client dialogue to be more robust
* prompt for default character creation on character card loads
* rename to model as to not conflict with pydantic
* narrate after dialogue strip dialogue generation unless enabled via new option
* starling and capybara-tess
* narrate dialogue context increased
* relabel tts agent to Voice, show agent label in status bar
* dont expect LLM to handle * and " - most of them are not stable / consistent enough with it
* starling template updated
* if allow dialogue in narration is disabled just assume the entire string is a narration
* reorganization the narrate after dialogue template
* fix more issues with time passage calculations
* move punkt download to agent init and silence
* improved RAG during conversation if AI selected is enabled in conversation agent
* prompt tweaks
* deepseek, chromomaid-storytelling
* relock
* narrate-after-dialogue prompt tweaks
* runpod status queries every 15 secs instead of 60
* default player character prompting when loading character card from talemate storage
* better chunking during split tts generation
* tweak narrate progress prompt
* improvements to ensure_dialogue_format and tests
* to pytest
* prep 0.15.0
* update packages
* dialogue cleanup fixes
* fix openai default model name
fix not being able to edit client due to name check
* free form analyst was using wrong system prompt causing gpt-4 to actually generate json responses
