MagicLens per AJ

MagicLens reads the text inside images on any web page and lays it back over the image, exactly where the original letters are, so you can select it, copy it, translate it or ask about it like normal page text.

How to use it
- Hover an image and click the lens button, or right-click it and choose Read text in image.
- Alt+Shift+L reads the main image on the page; Alt+Shift+K lets you pick any element.
- Drag to select, double-click a word, or copy everything with clean spaces and line breaks.

Local first
- Tesseract and PaddleOCR run entirely in your browser. Nothing leaves your device and no account is needed.
- English and PaddleOCR are built in; 29 more Tesseract languages download once when you turn them on.

Optional AI, your own key
- Translate the text in place, painted over the original like a printed page in your language.
- Explain, summarize or ask questions about the image.
- Hybrid mode lets a vision model correct the OCR while keeping exact word positions.
- Works with OpenAI, Claude, Gemini, OpenRouter, Groq, Mistral, xAI, Together, or a local model through Ollama or LM Studio. Images go only to the provider you choose, only when you ask.

Privacy
The developer collects nothing: no server, no account, no analytics, no ads. Full policy: https://axdsan.github.io/magiclens/privacy/

MagicLens is free and open source (Apache-2.0).