AI Avatar
Aria
Idle — Ready to Connect
Latency
STT
LLM
TTS
0:00Duration
🎙️ Voice Pipeline
STT EngineVibeVoice-ASR
TTS EngineChatterbox-Turbo
Sample Rate24kHz
HASM ClassArch B — SSR
Voice Profile
Aria ♀ Marcus ♂ Nova ◎
🧠 Brain Pipeline
LLMOllama — Qwen3:8B
PersonaCustomer Service
Context Window10 turns
Emotion DetectActive
ZD-LLM✓ Compliant
Shards MemoryPersistent
📋 20-Shot Consistency Gate
Avatar must complete 20-shot reference set via Character Engine before media creation.
0 / 20 shots verified
⚡ Latency Budgets Tier A
ms TTFA 0 breaches
ASR
LLM TTFT
TTS
TTFA
Idle Listen Think Speak
💬 Conversation 0 messages
Hello! I'm Aria, your AI assistant. How can I help you today?
Aria · just now
👤 Active Persona
Name
Role
Emotion😊 Friendly
Session
⚡ HASM v4 Pipeline
Data SourceArch B SSR
ExtractionSource-First
Bot ProtectionN/A (Local)
Benchmark>50 rec/sec
v5 PreviewStaged
🎬 MuseTalk Server Offline
Server URL
GPU
Modelv1.5
Avg Latency
Frames Generated0

👤 Persona Library

📋 20-Shot Consistency Verification

All avatar characters must complete a 20-shot reference set before media creation. This ensures visual consistency across all generated content.

⚡ Latency Budget Dashboard

📊 Budget Configuration
🎯 Tier Status
📈 Turn Latency History (last 20 turns)
🚨 Breach Log

🧩 Kappa Event Log

Append-only immutable event log. Frontend state = materialized view from log replay. Kappa Architecture

0 events