Home
Creators
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

PLUGIN UPDATE: Alive v54 (files) (Patreon)

Published:
2023-09-27 21:05:53
Imported:
2026-06
haven't archived this post yet. have a subscription? use the importer!

Content


CHANGELOG V54

  • FIX: characters, fixed error on loading character into the body of the same character
  • NEW: AI, added available word count for elevenlabs, requires changing tab if just enabling/disabling elevenlabs for now
  • NEW: AI, added support to get available elevenlabs AI models, option to pick which one to use; support for multilingual elevenlabs
  • FIX: UI removed posing long names, keep just numbers to better fit
  • FIX: UI fixed person atom still appearing sometimes in bottom bar even after turned off
  • NEW: UI, added option to start Alive with the UI not loaded in Plugin > Tools > Enable Alive UI for people that want just automatic behavior, UI can then be enabled when needed to setup some settings; You can also disable Alive Settings > Automatically Start to disable Alive auto-behaviour on al person atoms, that way you can just manually add plugins/AliveAlwaysOn.cslist on person atoms that you want to have Alive enabled in that scene, similar to the old person plugin (reload session plugin to see the effects on automatic behavior)
  • CHANGE: characters, added avatar photo change on save look similar to vam save look screenshot
  • NEW: characters, on load body load also the physical settings of that look (plugins, vam person settings), allows saving custom jaw settings, breasts & glutes physics, audio settings for each character; works with Bring to scene feature
  • NEW: effects, added person talking detection as effect; removing the talking effect will make the person stop saying whatever the model is saying
  • CHANGE: realism, turn off speech bubble notifications by default, redundant now with effects system
  • NEW: characters, added character load body + personality over the selected person atom feature
  • CHANGE: characters, changed character options popup to make it easier to use
  • CHANGE: UI, simplified voices test & list UI on services settings and character settings
  • FIX: UI, small fixes for UI alignment

CREDITS & INSTRUCTIONS


FILE

Files

Previews only

UPDATE: Alive v54

work in progress stuff for Alive (built with Virt-A-Mate): https://www.patreon.com/SPQR_aeternum credits AI voice: https://elevenlabs.io/ credits VAM: Roac - hair VRDollz - makeup, brows,earrings maru01 - dress rernat - eyelashes

Comments

Jonathon Barton

Is it possible to work in an "always on" option for TAT? I have it working in my own Speech to Speech solution. Here's the core of it... while True: try: # use SpeechRecognizer.Recognizer as sr and listen for speech on the microphone. If you don't hear anything, loop back and listen again. If you do hear speech, write the WAV data to an MP3 for transcription. with sr.Microphone(device_index=1, sample_rate=44100, chunk_size=512) as source: r.adjust_for_ambient_noise(source, duration=0.5) audio = r.listen(source, phrase_time_limit=60) with open(outfile, "wb") as f: # Write the audio data to an MP3 file f.write(audio.get_wav_data(convert_rate=44100, convert_width=2)) break except sr.RequestError as e: print("Request Error. Reason: ; {0}".format(e)) except sr.UnknownValueError: print("Silence detected. Say something or check your mic...") continue # If the MP3 file modified time changes, check to see if it's silent. If it's not, transcribe it. while transcribed==False: if os.path.exists(outfile): #audio_file = AudioSegment.from_file(outfile) modified = os.path.getmtime(outfile) if last_modified is None or modified > last_modified: last_modified = modified audio_nonsilent = detect_nonsilent(outfile, min_silence_len=500, silence_thresh=-10) if audio_nonsilent: print("Audio detected! Transcribing!") audio_file = open(outfile, "rb") try: response = openai.Audio.transcribe("whisper-1", audio_file) if response and 'text' in response: user_input = response['text'] user_input = str(user_input) transcribed = True time.sleep(5) # do something with the transcription print(user_input, "\n") # Sometimes I had issues with Whisper interpreting *almost* silence as "Thank you" - which I do see in Alive Chat from time to time... I changed "Thank you" responses to a random something else... if user_input.strip() == ". . . . ." or user_input.strip() == "Thank you": responses = ["*mumbles something you can't quite hear*", "*pauses in thought, searching for the right words...*", "Tell me more about that.", "*listens to you attentively as you continue..."] user_input = f"{random.choice(responses)}" else: return user_input else: print("Error: Invalid response from OpenAI API") if retry_count < max_retry: print("Retrying transcription...") retry_count += 1 time.sleep(5) else: break except openai.OpenAIError as error: # handle the exception print(f"OpenAI API Error: {error}") # handle arbitrary number of random responses responses = ["mumble something you can't quite hear.", "pauses in thought, searching for the right words..."] user_input = f"{human_character_name} {random.choice(responses)}" return user_input else: print("Audio is mostly silent. Skipping transcription.") else: print("Recording file does not exist. Waiting...") time.sleep(1) # And sometimes I'd get an error back from Whisper, or I'd get nothing back at all, even if there was good speech, sometimes Whisper just fails. So, intercept it and send *something* back anyway. if not transcribed: # handle arbitrary number of random responses responses = ["mumble something you can't quite hear.", "pauses in thought, searching for the right words..."] user_input = f"{human_character_name} {random.choice(responses)}" return user_input

SPQRAeternum
>>118546041

I considered it when making the script but I concluded that having it constantly record would be weird to a lot of people, and it could also create all sorts of fringe cases I don't have time to deal with

hashiba isuba

hello ! i'm sorry if this question is quite noob. I just wonder that can i disable some face morph ? I try to turn off anything that related to face/mouth in F2 UI e.g. lively mouth but it seems like some face morph still working e.g. surprise1 and some eyes. i love your plugin but with my models it looks really weird. Thanks !!

SPQRAeternum
>>118611107

Hi! No, unfortunately that's going to happen with some models depending on the morphs used. It works best with models that are close to neutral expressions by default. Lots of people add actual expressions on the look, like smiling or puckering lips, because vam is so static and that makes the models look better. But with Alive, the expressions would just stack up and there's no way that I found to prevent conflicts. There is the same problem with eyes, blinking going sometimes over the board if the model has eyes morphs that already close them. If you can DM me an example model I can try to adjust the morphs to make it slightly better. But no promises. The goal is to make it work with neutral models. If I tune down too much for some models that have expression morphs already, expressions will become barely noticeable for neutral models.