AI Models

18 models

Download and manage Whisper and LLM models for transcription and summarization.

Overview

EdgeNote AI uses two types of AI models that run locally on your device:

Whisper Models
5 models

OpenAI's Whisper models for speech-to-text transcription. Choose based on accuracy needs and language support.

LLM Models
13 models

Large Language Models for summarization and insight extraction. Choose based on quality needs and available RAM.

Whisper Models (Transcription)

Models for converting speech to text.

ModelSizeRAM RequiredLanguagesNotes
Whisper Tiny
77 MB1 GBEnglishUltra-fast, lower accuracy
Whisper Small
488 MB2 GBEnglishFast with good accuracy
Whisper MediumRecommended
1.5 GB4 GBEnglishExcellent accuracy
Whisper Large V3 Turbo
1.6 GB4 GB99 languagesFast + multilingual
Whisper Large V3
3.1 GB6 GB99 languagesMaximum accuracy

LLM Models (Summarization)

Models for generating summaries and extracting insights.

ModelSizeRAM RequiredNotes
Qwen 3 1.7B
1.4 GB2 GBLightweight and fast
DeepSeek R1 1.5B
1.1 GB2 GBFast with chain-of-thought
Gemma 2 2B
1.6 GB4 GBCompact and efficient
Phi-3.5 Mini 3.8B
2.4 GB4 GBBalanced performance
Llama 3.2 3B
2.0 GB4 GBFast and reliable
Qwen 3 4BRecommended
2.6 GB4 GBBest for most desktops
Mistral 7B v0.3
4.4 GB12 GBFast inference
DeepSeek R1 7B
4.7 GB12 GBExcellent reasoning
Llama 3.1 8B
4.9 GB12 GBGeneral purpose
Qwen 3 8B
5.2 GB16 GBSuperior quality with thinking
Gemma 2 9B
5.8 GB8 GBStrong performance
DeepSeek R1 14B
9.0 GB16 GBSuperior reasoning
Phi-4 14B
9.1 GB16 GBState-of-the-art

Downloading Models

Models are downloaded on first use or can be pre-downloaded from Settings:

1

Open Settings

Go to Settings > Transcription or Settings > Summarization.

2

Select a Model

Choose a model from the dropdown. Models show size and RAM requirements.

3

Download

Click download. Progress is shown in the interface. Models are stored locally and don't need to be re-downloaded.

Model Download Progress
Downloading a model shows progress and estimated time

Custom Model Downloads

Advanced users can download custom GGUF models from Hugging Face:

Adding Custom LLM Models

  1. Find a GGUF model on Hugging Face (e.g., TheBloke's quantized models)
  2. Copy the model URL (must end in .gguf)
  3. Go to Settings > Summarization > Custom Models
  4. Paste the URL and click Add
  5. The model will download and appear in your model list
Custom Model Management
Adding and managing custom GGUF models

Model Management

View Installed

See all downloaded models with their size and location on disk.

Delete Models

Remove unused models to free up disk space. Can be re-downloaded anytime.

Switch Models

Change the active model at any time. New recordings will use the selected model.

Check Updates

EdgeNote AI will notify you when newer model versions are available.

Choosing the Right Models

See the Hardware Requirements page for detailed recommendations based on your available RAM.

Use CaseWhisperLLM
Quick notes (8GB RAM)Whisper SmallQwen 3 1.7B
Most users (16GB RAM)Whisper MediumQwen 3 4B
Multilingual (16GB RAM)Large V3 TurboQwen 3 4B
Complex analysis (24GB+ RAM)Whisper Large V3DeepSeek R1 7B
Maximum quality (32GB+ RAM)Whisper Large V3Phi-4 14B