メインコンテンツまでスキップ

Release notes

ACUAH (ver.3.0.0-)​

ver.3.5.0​

[9-28-2024]

Important

Due to the Gemini API support, the LLM prompt needs to be reconfigured. When updating from ver. 3.4.x, the existing prompt settings will be taken over, but may not work properly. We apologize for the inconvenience, but please initialize and save the existing settings by "Clear" in the AI(LLM) prompt settings screen, and then reconfigure the settings.

  • Open the AI(LLM) prompt settings screen.
  • Select an existing prompt, and note the API key and the string of the prompt.
  • Initialize the relevant prompt settings by clicking "Clear" button, and save it by clicking "Save" button.
  • Set the obtained API key and the prompt string again, and click "Save" button to save the settings.
  • Select the appropriate prompt on the character setting screen and tap the "OK" button.
  • Support Gemini API
  • Function changed
    • Remove the "Function calling" setting.
      • Function calling is "disable" for conversation type "LLM"
      • Function calling is "enable" for conversation type "Scenario/LLM"
  • Bug fixes
    • The Azure TTS ApiKey setting is incorrect in the Voice settings.
    • Lip-sync doesn't work properly with some character models.
    • VRoidHub character model doesn't desplay when it's cache data is corrupted.

ver.3.4.1​

[8-28-2024]

  • Bug fixes (for Android)
    • Unable to add music files to the initial music list.(musiclist0)

ver.3.4.0​

[8-27-2024]

  • Added functions
    • Phone dial-out function
    • Smartphone on-screen keyboard input button
    • Start voice recognition with Bluetooth speaker play/pause button and headset hook button.
  • Update SDK and libraries to the following.
    • UniVRM 0.124.2
  • Changed default model of ChatGPT to gpt-4o-mini.
  • Other minor fixes

ver.3.3.1​

[5-18-2024]

  • Bug fix
    • When the conversation type is "dialogue scenario only," there are cases in which the character does not learn properly.

ver.3.3.0​

[5-10-2024]

  • Added functions
    • Learning function using return value of ChatGPT function calling arguments.
    • Simple delete function for learned data.
  • Other minor fixes

ver.3.2.0​

[4-27-2024]

  • Added functions
    • Real world screen for the character to view (web camera)
    • Photo
    • Photo album
    • QR code reader
  • Bug fixes
    • May not handle ChatGPT response properly.
    • The character assistant didn't wake up (appear) from sleeping before waking the user up.
  • Other minor fixes

ver.3.1.0​

[4-4-2024]

  • Added functions
    • Execute function with ChatGPT Function calling.
    • Paste the text from clipboard to note.(Notepad function)
    • Copy the text of shown memo to clipboard.(Notepad function)
  • Bug fixes
    • Conversation type settings were not saved properly.

ver.3.0.0​

[3-24-2024]

  • Added function to set multiple ChatGPT prompts
  • Added function to set multiple voice data sets
  • VRMA (VRMAnimation) loading support
  • Updated SDKs and libraries to the following
    • VRM 0.120.0
    • VRoidSDK 0.5.1
  • Function changed
    • Changed backend server from NCMB to ParseServer
    • Removed CeVIO voice data to reviewing of rights.
    • Changed so that the character does not speak when the only conversation type is ChatGPT.
    • Revised the position and display of buttons.
    • Fixed some motions.
    • (below for Android)
      • Changed icon design
      • Changed lip-sync to SALSA LipSync Suite
      • Supported multiple file selection when selecting music files
      • Supported multiple file selection when loading motion data
      • Changed file extension of notepad file (notexxxx) to ".txt"
  • Bug fixes
    • ChatGPT prompt edit does not support line breaks.
    • Lip-sync does not work with some character models.
    • (below for Android)
      • Misskey authentication sometimes failed.
  • Other minor fixes

ACUAH (ver.2.0.0-)​

ver.2.0.2​

[3-3-2024]

  • Bug fixes (for iPhone)
    • Notepad image is corrupted.

ver.2.0.1​

[2-13-2024]

  • Released ACUAH for iPhone
    • Almost the same specifications as Android version ver.2.0.0.
    • Some functions of the Android version are not available.
      • Bluetooth devices (multimedia remote control, keyboard)
      • Misskey integration

ver.2.0.0​

[7-14-2023]

  • Support VRM1.0 format
  • Updated each SDK and libralies to the following
    • VRM 0.112.0
    • VRoidSDK 0.2.2
    • EVMC4U ExternalReceiverPack_v3.9c_VRM1 (deprecated version)
  • Function changed
    • Changed standard screen image quality from "Medium" to "High"
    • Changed the default value of Bloom from 1.0 to 0.0 (to reduce power consumption)
  • Fixed button placement
  • Other minor bug fixes

ACUAH (ver.1.0.0-)​

ver.1.7.0​

[6-17-2023]

  • Added functions
    • Added support for ChatGPT model version change.
    • Added an action to launch the browser.
    • Added support for playing external files as sound effects and background music (ACUAH editor required)
  • Function changed
    • Changed recommended device to Android13.
    • Changed Twitter integration to Misskey (misskey.acuah.info) integration.
  • Bug fixes
    • ChatGPT response may not work correctly after ChatGPT process is interrupted.
    • Alarm sound does not play correctly on Android13.
    • Bluetooth ON/OFF does not work on Android13.
    • Some character models are not displayed correctly
    • Voice recognition does not work properly and sometimes gets stuck.
    • Sleep voice does not play in sleep mode.
  • Other minor bug fixes

ver.1.6.1​

[5-15-2023]

  • Modification of the following functions
    • Revised ChatGPT response
    • You can change character size, rotation, position (up/down) by touch gesture when the character is not shown.
  • Bugs fixes
    • Lip sync does not work properly when playing VoiceAPI for some character models.

ver.1.6.0​

[5-7-2023]

  • Added a few motions (Idle state, Look at left/right etc.)
  • Added function
    • You can change character size, rotation, position (up/down) by touch gesture.
  • Other minor bug fixes

ver.1.5.0​

[4-15-2023]

  • ChatGPT API/Azure Text-To-Speech support (require your API key)
  • Bug fixes
    • fixed problems with the application not working properly in some OS languages (System>Languages setting) (slow operation, characters' facial expressions significantly broken, etc.)
  • Other minor bug fixes

ver.1.4.2​

[4-5-2023]

  • Bug fixes
    • Lip-sync does not work properly on some character models.
  • Other minor bug fixes

ver.1.4.1​

[3-31-2023]

  • Modification of the following functions
    • Revised VOICEVOX server access method.
  • Other minor bug fixes

ver.1.4.0​

[3-21-2023]

  • ChatGPT API/VOICEVOX support (Sorry, Japanese only) (test implementation)
  • Function changed
    • Revision of dialogue scenario format
    • Modification of weather forecast and rock-paper-scissors scenarios
    • Create idle response scenarios (supported editing and replacement of all voice data in ACUAH editor)
  • Bug fixes
    • Continuous playback of audio files does not work properly
    • Dialogue scenario item action "WAIT" does not work correctly.
    • Models without Hand/Hips bones are not displayed with an error.

ver.1.3.1​

[2-10-2023]

  • Removed function
    • Sony Corporation mocopi™ (mocopi Receiver Plugin) support

ver.1.3.0​

[1-23-2023]

  • Added function
    • Sony Corporation mocopi™ (mocopi Receiver Plugin) support (This is a test implementation. Please note that this function may be removed in the future due to the license agreement with Sony Corporation.)
  • Other minor bug fixes

ver.1.2.1​

[1-18-2023]

  • Bug fixes
    • Character model change method setting may not be saved properly
  • Other minor bug fixes

ver.1.2.0​

[1-9-2023]

  • Added a few motions
  • Modification and additions of the following functions
    • Modified to save playback method (shuffle, repeat) and reverb zone settings for each music list
    • Twitter integration - modified so that ACUAH learns if the first sentence tweeted by ACUAH is liked
  • Bug fixes
    • Effects at the bottom of the screen were not reflected when the character position was moved upward in the camera settings
    • Wrong settings for right and left hands on some character models
    • Hair and other objects do not swing in some character models
    • Some BVH files may not be loaded properly
    • Settings may be reset after setting a LocalVRM file to a character
  • Other minor bug fixes

ver.1.1.2​

[12-15-2022]

  • Bug fixes
    • Character models do not display properly under certain conditions.
  • Other minor bug fixes

ver.1.1.1​

[12-7-2022]

  • Bug fixes
    • The position of the chair is incorrect.

ver.1.1.0​

[12-2-2022]

  • Modification and additions of the following functions
    • Added to select alarm ringtone from Android OS alarm settings (no alarm sound is also available)
    • Added to adjust the vertical position of the model when setting the camera (not adjustable when using the false-hologram smartphone stand)
    • Added some objects for character models to hold (smartphone, notepad, pencil, eraser)
    • Revised buttons (UI) on each screen, and changed so that they are automatically hidden on the character display screen
  • Bug fixes
    • Overall review of application and asset file update process
  • Other minor bug fixes

ver.1.0.1​

[11-10-2022]

  • Modification of the following functions
    • Changed the default value of the usage setting of the false-hologram smartphone stand to disable
  • Bug fixes
    • Char2 to char4 cannot reset the character
    • App may crash when selecting LocalVRM data

ver.1.0.0​

[10-17-2022]

  • Changes to default character model (ACUAH)
  • Removed DMM VR Connect β linkage
  • Support in-application updates on Google Play
  • Update each SDK to the following
    • VRoidSDK 0.1.1
    • UniVRM 0.95.1
    • uOSC v2.1.0
    • EVMC4U v3.9a
  • Changed color space from Gamma to Linear
  • Added effect (Bloom) setting
  • Added motion data (BVH) loading support
  • Added character naming settings (avatar OK models only)
  • Added emotion value parameters
  • Functions changed
    • Changed VRoidHub linkage requirement to optional
    • Revised overall layout of each UI, etc.
    • Added selection to use hashtag (#ACUAH) when tweeting on Twitter
    • Removed the character hide button and added UI (display orb)
    • Changed voice recognition start from button to character tap
    • Changed VRM file selection from any folder (Android file picker) instead of specified local folder
  • Bug fixes
    • Japanese cannot be input with Bluetooth keyboard
    • VMC protocol receiving may not work
    • IFTTT Webhooks trigger may not be sent properly
    • Dance may end by itself unintendly when dance is executed continuously
  • Other minor bug fixes

ACUAH（ver.0.9.25～）​

ver.0.9.29​

[7-16-2022]

  • ACUAH user account registration and deletion (optional)
    • Server backup/restore of configuration information
    • Server upload of training data for dialogue scenarios
    • Server upload of tweet learned data for Twitter integration
  • Implementation of dialogue scenarios (play rock-paper-scissors)
  • Modification of the following functions
    • Each configuration screens
    • Download function from server
    • Unified English notation for Android OS system languages other than Japanese and English
    • Twitter tweets
    • Addition of option to not reflect affection value in character’s facial expression
  • Bug fixes
    • Bluetooth ON/OFF does not respond on Android12
    • Crash after 80 minutes of application startup on Android12
    • Motion may not play after wake up alarm
  • Other minor bug fixes

ver.0.9.28​

[4-2-2022]

  • Changed dialogue scenario format
  • Added posing motion for screenshots
  • Changed camera setting scenario and added gyro-sensor linkage
  • Changed sound effect data
  • Added background and lighting setting functions
    • WebCam, for chroma key (blue/green/red)
  • Added Twitter reply function (test implementation)
  • Changed UniVRM version 0.62 to 0.66
  • Temporarily hides alarm time display while playing music
  • Other minor bug fixes

ver.0.9.27​

[2-4-2022]

  • Added idle motion data (standing pose, sitting pose)
  • Added new motions in playing music
  • Added music playback options
    • Multiple list setting
    • Repeat play
    • Reverb zone setting
  • Added character model’s facial expression parameter adjustment
  • Change motion during timer (TypeB only)
  • Other minor functions and bug fixes
    • Temporarily hide alarm time while dancing
    • Fix some shaders were not loaded properly
    • Fix lighting setting

ver.0.9.26​

[12-12-2021]

  • Added motion data
  • Added a function to select the type of motion (TypeA/B).
  • Added some motions and effects to Unite In The Sky (© UTJ/UCL).
  • Added Android notification function (test implementation)
  • Added Twitter integration function (test implementation)
  • Fixes other minor bug.

ver.0.9.25​

[10-9-2021]

  • Change application icon.
  • Overall reconstruction of animation and line of sight processing.
  • Support VMC Protocol Marionette (EVMC4U)
  • Support loading VRM file from smartphone local folder.
  • Support ACUAH editor.

ACUAH β (～ver.0.9.24)​

ver.0.9.24​

[7-2-2021]

  • Changed effects for summoning, invocation.
  • Added effects when playing music.
  • Added lip-sync animation in English voice. (quality Medium or higher)
  • Added information window about this application.
  • Fixed the following bugs.
    • Characters do not reappear after sleep mode.
    • When a character recovers from sleep mode, the motion and summon particle are doubled.
    • Data download function from backend server.
  • Other minor bug fixes.

ver.0.9.23​

[5-13-2021]

  • English support.
  • The application dialogue scenarios are changed with Android OS languages. (System > Languages & input > Languages)