VOICE COMPANION RUNTIME

Noema

STATE

Desktop AI companion for voice, memory, and tools.

PERSONA BOOTTEXTURE ASSEMBLY

Provider matrixRuntime boundary

LLM corereasoning + tool planning

Ollama

OpenAI

Qwen

Gemini

Claude

DeepSeek

Groq

ASR inputspeech recognition

OpenAI

Qwen

AssemblyAI

Google Cloud

Azure AI

TTS outputvoice synthesis

ElevenLabs

OpenAI

Azure AI

Minimax

Fish Audio

Terminal diagnostics

$ noema doctor --providers
Runtime Boundary : active
Model Bus        : connected
Providers        : 14 / 14 reachable
Latency (p95)    : 128ms
Local Runtime    : Ollama

provider                  type       status    p95 latency
ollama/llama-3.1-8b       LLM        ready     118ms
openai/gpt-4o-mini       LLM        ready     112ms
qwen/qwen2.5-7b          LLM        ready      96ms
qwen/qwen-audio          ASR        ready     109ms
assemblyai/universal      ASR        ready     104ms
google-cloud/speech       ASR        ready     121ms
elevenlabs/voice          TTS        ready     127ms
minimax/speech            TTS        ready     134ms
fish-audio/speech         TTS        ready     127ms

Audio visualizer

N O E M A
RuntimeSection 3 / 3
ExtendNR-7 ○
Build mode
Local

Build the companion loop.

Extend Noema with plugins, tools, memory, and local workflows.

Read docs View GitHub Write a plugin