tant/yuna
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
Làm thử AI phỏng vấn xem sao
2 commits 1 branch 0 tags 250 KiB
  • TypeScript 73%
  • Python 15.2%
  • Shell 5.8%
  • CSS 5.4%
  • Dockerfile 0.3%
  • Other 0.3%
Find a file
Exact
tant f24a1f205d Add multi-language support and rebuild frontend with Next.js 16 
Backend changes:
- Add language-specific system prompts for LLM (vi/en/auto)
- Add TTS voice switching based on language selection
- Add auto language detection from text for TTS in auto mode
- Fix health check attribute after TTS refactor
- Update WebSocket handler to propagate language to all services

Frontend rebuild:
- Migrate to Next.js 16.1.3 with React 19
- Add 3D avatar with React Three Fiber (professional 3/4 view)
- Implement VAD-based voice detection
- Add language selection in setup panel (Vietnamese/English/Auto)
- Add real-time processing indicators
- Build reusable UI component library

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-01-17 12:12:20 +07:00
backend Add multi-language support and rebuild frontend with Next.js 16 2026-01-17 12:12:20 +07:00
docs Initial commit: Local Voice AI Interview backend 2026-01-17 08:18:58 +07:00
frontend Add multi-language support and rebuild frontend with Next.js 16 2026-01-17 12:12:20 +07:00
models Initial commit: Local Voice AI Interview backend 2026-01-17 08:18:58 +07:00
.env.example Initial commit: Local Voice AI Interview backend 2026-01-17 08:18:58 +07:00
.gitignore Initial commit: Local Voice AI Interview backend 2026-01-17 08:18:58 +07:00
CLAUDE.md Add multi-language support and rebuild frontend with Next.js 16 2026-01-17 12:12:20 +07:00
docker-compose.yml Initial commit: Local Voice AI Interview backend 2026-01-17 08:18:58 +07:00
README.md Initial commit: Local Voice AI Interview backend 2026-01-17 08:18:58 +07:00
select-models.sh Initial commit: Local Voice AI Interview backend 2026-01-17 08:18:58 +07:00

README.md

Yuna - Local Voice AI Interview

100% offline voice AI interview practice with 3D avatar.

Tech Stack

  • Backend: FastAPI + Pipecat
  • STT: faster-whisper (local)
  • LLM: Ollama (qwen3:4b)
  • TTS: Piper (local)
  • Frontend: Next.js + TalkingHead.js (planned)

Quick Start

# Backend
cd backend
uv sync
source .venv/bin/activate
uvicorn src.main:app --reload --port 8765

Requirements

  • Python 3.12
  • Ollama with qwen3:4b model
  • Piper voice models in models/piper/