Releases: uezo/ChatdollKit
v0.8.12
v0.8.11
🤖 Support Server-side Agent Framework Collaboration
Offloads AI agent logic to the server—boosting front-end maintainability—while letting you plug in frameworks like AutoGen (and any other agent SDK) for unlimited capability expansion.
- Add support for AIAvatarKit as an AI agent backend service
- Support
inputsparameter for Dify - Fix bug where API Key authorization doesn't work in WebGL
- Allow null for SystemPromptParams for AIAvatarKit
🌐 WebGL Improvements
Upgraded mic capture to modern AudioWorkletNode for lower latency and reliability; stabilized mute/unmute handling; improved error handling to immediately surface HTTP errors and prevent hangs; fixed API-key authorization in WebGL builds.
- Switch WebGLMicrophone implementation from ScriptProcessor to AudioWorklet
- Prevent processing of muted user's speech after unmuting on WebGL
- Return HTTP errors immediately to avoid AI character hanging
🍩 Other updates
- Remove CommandRService
- Add option to include Wave header in SpeechGatewaySpeechSynthesizer
- Add channel info to ChatMemory integration extension
- Enable downsampling of microphone input for speech recognition
By the way, this release was prepared while enjoying a picnic at Jonanjima Seaside Park 🏕️🌳
Full Changelog: 0.8.10...v0.8.11
v0.8.10
🌎 Dynamic Multi-Language Switching
- Support dynamic multi language speech synthesizing #414
- Enable Multi-Language Support for Speech Recognition #415
- Add support for SpeechGateway in SpeechSynthesizer #416
🔖 Long-Term Memory
- Add ContextId for conversation-level identification #418
- Add request info to the params of
OnStreamingEnd#419 - Add support for Long-Term Memory #420
🍩 Other Updates
- Support IsAzure option of ChatGpt LLM by AITuber controller by @buchizo in #413
- Enable echo cancellation and noise suppression for WebGL #417
- Prevent WebGL build error #421
- Improve error handling in HTTP access #422
- Fix bug where setting Nijivoice duration fails #423
- Small changes and update README for v0.8.10 #424
Thank you so much for your contribution, @buchizo san!🥰🥰🥰
Full Changelog: 0.8.9...0.8.10
Contributors
Assets 6
v0.8.9
✨ Support NijiVoice as a Speech Synthesizer
🍩 Other changes
- Improve dialog processing #410
- Fix bug where DialogProcessor fails on before processing LLM stream #411
- Update for v0.8.9 #412
Full Changelog: 0.8.8.1...0.8.9
v0.8.8.1
💪Support Dify as a backend for AITuber
Seamlessly integrate with any LLM while empowering AITubers with agentic capabilities, blending advanced knowledge and functionality for highly efficient and scalable operations!
Full Changelog: 0.8.8...0.8.8.1
v0.8.8
v0.8.7
✨More Features For AITuber and Update Demo✨
- Support start/stop SocketServer from external components #395
- Add dummy components for the use case that doesn't use microphone #396
- Update demo for v0.8.7 #397
- Update README and small changes for v0.8.7 #399
🐛 Bug fix
- Fix bug where WebGL build fails #398
Full Changelog: 0.8.6...0.8.7
v0.8.7 beta OOTB app
Out-of-the-box application for AITuber use case.
Use ChatdollKit AITuber Controller to control.
chatdollkit_aituber_ootb_fhd.mp4
v0.8.6
🎛️ Support VOICEVOX and AivisSpeech inline style
Enables dynamic and autonomous switching of voice styles to enrich character expression and adapt to emotional nuances.
- Support applying inline style for VOICEVOX #390
🥰 Improve VRM runtime loading
Allows seamless and error-free switching of 3D models at runtime, ensuring a smoother user experience.
- Resolve runtime VRM model loading errors and refactor configuration process #391
🍩 Other changes
- Enhance AITuber Demo with New REST API Features #392
Full Changelog: 0.8.5...0.8.6
v0.8.5
🎓 Chain of Thought Prompting
Chain of Thought (CoT) prompting is a technique to enhance AI performance. For more information about CoT and examples of prompts, see https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/chain-of-thought .
ChatdollKit supports Chain of Thought by excluding sentences wrapped in ~ tags from speech synthesis.
You can customize the tag by setting a preferred word (e.g., "reason") as the ThinkTag in the inspector of LLMContentProcessor.
- Enable Chain of Thought Prompting #387
🍩 Other updates
Full Changelog: 0.8.4.1...0.8.5