{#if visible}

{ if (e.key === 'Escape') handleCancel(); }}>

Settings

{#if activeTab === 'transcription'}

Whisper Model

Device

Language (blank = auto-detect)

{:else if activeTab === 'speakers'}

HuggingFace Token

Why is this needed?

Speaker detection uses the pyannote.audio model, which is hosted on HuggingFace and requires accepting a license agreement.

How to get a token:

Create a free account at huggingface.co
Go to huggingface.co/pyannote/speaker-diarization-3.1 and accept the license
Go to huggingface.co/settings/tokens and create a token with read access
Paste the token above and click Save

The model will be downloaded automatically on first use (~100 MB).

Skip speaker detection (faster, no speaker labels)

{:else if activeTab === 'ai'}

AI Provider

{#if localSettings.ai_provider === 'openai'}

OpenAI API Key

Model

{:else if localSettings.ai_provider === 'anthropic'}

Anthropic API Key

Model

{:else if localSettings.ai_provider === 'litellm'}

Model

{/if} {:else}

llama-server Binary Path

GGUF Model Path

Place GGUF model files in ~/.voicetonotes/models/ for auto-detection. The local AI server uses the OpenAI-compatible API from llama.cpp.

{/if}

{/if}