LensLore - Browser Vision & Speech

⚙️ Settings

Changes are saved automatically

Device Preference

Choose how to run the AI model

HuggingFace Mirror

Enable mirror for faster downloads in China (experimental)

If download fails, you can download files manually (e.g. using IDM/Thunder) and import them here.
Download Vision Files | Download TTS Files | Download ASR Files

Vision Model (SmolVLM) Checking...

Required: config.json, tokenizer.json, onnx/model_fp16.onnx (or model.onnx)

TTS Model (Kokoro) Checking...

Required: config.json, model.onnx

ASR Model (Whisper) Checking...

Required: config.json, encoder_model.onnx, decoder_model_merged.onnx, vocab.json

Click "Save & Apply" after selecting

Default

Describe this image in 2-3 concise sentences.

Brief

Describe this image briefly.

Detailed

Describe what you see in plain text, without any formatting or markdown. Be natural and conversational.

Single Sentence

Describe in one sentence.

Main Subject

What's the main subject?

Custom

Write your own prompt below ↓

✅ 100% Offline Processing

All AI models run entirely in your browser. Your images, audio, and text never leave your device.

📥 One-time Model Download

Models are downloaded from HuggingFace. Once downloaded, models are cached for offline use.

🔒 No Data Upload

We don't collect, store, or upload any of your data. Everything stays on your device.

Version 1.0.11