Model Management Commands

Local STT Models

download_local_stt_model

Downloads a local speech-to-text model from Hugging Face.

await invoke('download_local_stt_model', {
  model: 'nvidia/parakeet-tdt_ctc-110m'
});

model

string

required

Model identifier (e.g., nvidia/parakeet-tdt_ctc-110m, usefulsensors/moonshine-base)

Response:

model

string

Canonical model name

provider

string

Model provider (parakeet, whisper, moonshine, etc.)

method

string

Download method used

localPath

string

Local filesystem path where model is stored

details

string

Human-readable status message

warmup_local_stt_model

Preloads a local STT model into memory for faster inference.

await invoke('warmup_local_stt_model', {
  model: 'nvidia/parakeet-tdt_ctc-110m'
});

model

string

required

Model identifier to warm up

Response:

model

string

Model name

provider

string

Model provider

warmed

boolean

Whether warmup succeeded

details

string

Status details

deactivate_local_stt_model

Unloads a local STT model from memory.

await invoke('deactivate_local_stt_model', {
  model: 'nvidia/parakeet-tdt_ctc-110m'  // Optional
});

model

string

Model to deactivate (omit to deactivate all)

Response:

model

string

Deactivated model name

provider

string

Model provider

deactivated

boolean

Whether deactivation succeeded

details

string

Status details

get_local_stt_hardware_advice

Analyzes system hardware and recommends appropriate STT models.

await invoke('get_local_stt_hardware_advice', {
  selectedModel: 'nvidia/parakeet-tdt_ctc-110m'  // Optional
});

selectedModel

string

Currently selected model to validate against hardware

Response:

cpuName

string

CPU model name

logicalCores

number

Number of logical CPU cores

totalRamGb

number

Total system RAM in gigabytes

nvidiaGpuDetected

boolean

Whether NVIDIA GPU is available

gpuName

string

GPU model name

gpuVramGb

number

GPU VRAM in gigabytes

performanceTier

string

Hardware tier: high, medium, low

slasshySuggestionModel

string

Recommended model for this hardware

suggestedModels

string[]

List of recommended models

cautionModels

string[]

Models that may be too demanding

selectedModelWarning

string

Warning message if selected model is unsuitable

details

string

Additional hardware details

delete_local_stt_model

Deletes a downloaded local STT model.

await invoke('delete_local_stt_model', {
  model: 'nvidia/parakeet-tdt_ctc-110m'
});

model

string

required

Model identifier to delete

Response:

model

string

Deleted model name

repoId

string

Hugging Face repository ID

removed

boolean

Whether deletion succeeded

localPath

string

Path that was deleted

details

string

Deletion status

Ollama Models

pull_ollama_model

Downloads a model from Ollama registry.

await invoke('pull_ollama_model', {
  baseUrl: 'http://localhost:11434',  // Optional
  model: 'llama3.2:3b'
});

model

string

required

Ollama model identifier (e.g., llama3.2:3b, mistral:latest)

baseUrl

string

Ollama API base URL (defaults to http://localhost:11434)

Response:

baseUrl

string

Ollama base URL used

model

string

Model name

boolean

Whether pull succeeded

status

string

Pull status message

get_ollama_status

Checks if Ollama is installed and running.

await invoke('get_ollama_status', {
  baseUrl: 'http://localhost:11434'  // Optional
});

baseUrl

string

Ollama base URL to check

Response:

installed

boolean

Whether Ollama is installed

running

boolean

Whether Ollama service is reachable

version

string

Ollama version string

details

string

Additional status information

fetch_provider_models

Fetches available models from an OpenAI-compatible provider.

await invoke('fetch_provider_models', {
  apiKey: 'sk-...',
  apiBaseUrl: 'https://api.openai.com/v1'  // Optional
});

apiKey

string

required

Provider API key

apiBaseUrl

string

API base URL (defaults to OpenAI)

Response:

baseUrl

string

Base URL used

models

string[]

List of available model IDs

TTS Voice Management

ensure_voice_model

Downloads and installs a Piper TTS voice model.

await invoke('ensure_voice_model', {
  // Internal command - automatically handled by setup_assistant_runtime
});

Response:

modelPath

string

Path to installed voice model

configPath

string

Path to voice config file

clone_coqui_voice

Clones a voice from audio sample using Coqui TTS.

await invoke('clone_coqui_voice', {
  pythonPath: '/path/to/python',  // Optional
  modelName: 'tts_models/multilingual/multi-dataset/xtts_v2',  // Optional
  language: 'en',  // Optional
  speakerId: 'my_voice',
  audioBase64: voiceSampleBase64,
  fileName: 'sample.wav',  // Optional
  useGpu: true  // Optional
});

speakerId

string

required

Unique ID for the cloned voice

audioBase64

string

required

Base64-encoded audio sample (5-30 seconds recommended)

pythonPath

string

Custom Python executable path

modelName

string

Coqui TTS model to use

language

string

Target language code

fileName

string

Original audio filename

useGpu

boolean

Enable GPU acceleration

Response:

speakerId

string

Created voice ID

durationSeconds

number

Audio sample duration

voiceDir

string

Voice storage directory

voices

string[]

List of all available voices

previewAudioBase64

string

Base64-encoded preview audio

Example Usage

Local STT
Ollama
Voice Cloning

// Download and warm up a local STT model
await invoke('download_local_stt_model', {
  model: 'nvidia/parakeet-tdt_ctc-110m'
});

await invoke('warmup_local_stt_model', {
  model: 'nvidia/parakeet-tdt_ctc-110m'
});

// Check Ollama status and pull a model
const status = await invoke('get_ollama_status', {});

if (status.running) {
  await invoke('pull_ollama_model', {
    model: 'llama3.2:3b'
  });
}

// Clone a voice from audio
const result = await invoke('clone_coqui_voice', {
  speakerId: 'john_voice',
  audioBase64: sampleAudio,
  language: 'en',
  useGpu: true
});

console.log('Cloned voice:', result.speakerId);

Architecture

Commands

Types

Model Management Commands

Local STT Models

download_local_stt_model

warmup_local_stt_model

deactivate_local_stt_model

get_local_stt_hardware_advice

delete_local_stt_model

Ollama Models

pull_ollama_model

get_ollama_status

fetch_provider_models

TTS Voice Management

ensure_voice_model

clone_coqui_voice

Example Usage

Build docs developers (and LLMs) love

Architecture

Commands

Types

​Local STT Models

​download_local_stt_model

​warmup_local_stt_model

​deactivate_local_stt_model

​get_local_stt_hardware_advice

​delete_local_stt_model

​Ollama Models

​pull_ollama_model

​get_ollama_status

​fetch_provider_models

​TTS Voice Management

​ensure_voice_model

​clone_coqui_voice

​Example Usage

Build docs developers (and LLMs) love

Local STT Models

download_local_stt_model

warmup_local_stt_model

deactivate_local_stt_model

get_local_stt_hardware_advice

delete_local_stt_model

Ollama Models

pull_ollama_model

get_ollama_status

fetch_provider_models

TTS Voice Management

ensure_voice_model

clone_coqui_voice

Example Usage