Command Palette

Search for a command to run...

Live demo
You are browsing real pipeline data captured from a live 5-Domia home mesh — latencies, delegation journeys, memories and emotions are genuine. Voices are synthetic. The demo is read-only.
Domias

Sous

Offline·Thin client
DOMIA_SOUS192.168.0.8:3111·Config-active·Last interaction 4 hours ago
ConfigureTalk to this Domia
Performance

First audio p50

1.69 s

p95 2.78 s

Total p50

11.27 s

p95 21.60 s

Interactions

10

Execution

0 local

10 delegated

STTstreaming-zipformer-en9×145 ms
LLMllama3.1:8b10×1.77 s
TTSKOKORO9×2.07 s
Stage waterfall (avg)
STT145 msLLM1.72 sTTS2.07 s

First audio at 1.89 s — the user hears Domia speak 2.05 s before the full 3.94 s pipeline finishes, thanks to per-sentence LLM→TTS pipelining.

Persona
Character
Sous
Personality
PLAYFUL
Profession
CHEF
Style
ENTHUSIASTIC
Perceived age
ADULT
Knowledge
EXPERT
Relationship
HELPER
Role mode
ACTIVE
Language
en

Interests

seasonal producefermentationpastry

Hobbies

inventing midnight snacksjudging other people's knife skills

Skills

recipesmeal planningwine pairing
Engines
Read-only · via DB

LLM

Engine
OLLAMA
Model
llama3.1:8b
Temperature
0.7
Context window
4096
Max tokens
80
Concurrency
3

TTS

Engine
KOKORO
Voice
af_bella
Language
en
Speed
1

STT

Engine
ZIPFORMER
Model
streaming-zipformer-en
Language
en

Wake word

Engine
KWS
Keyword
computer
Framework
onnx
Modules
Emotion engine
Emotion capture
Memory engine
Fact capture
Fact recall
Collective mind
Narrative engine
Identity engine
Remote access
Reflect when idle
Reflection concurrency
2
Mood
Capabilities
Read-only · via DB
Wake word
Record
Speech to text
Intent detection
Intent execution
Prompt generation
Language model
Text to speech
Playback

Delegations

None — runs everything locally.

Skills
Read-only · via DB

No skill providers configured.

Recent interactions