# OCR Extractor plugin for Obsidian OCR Extractor is an [[Obsidian]] community plugin by Johnathan Ritzi that extracts text from attachments embedded in your notes (images, PDFs, documents) and writes it as [[Markdown]] in a collapsible callout right below the embed. That's the approach I recommend: text that isn't in a Markdown file isn't reliably findable. Once the text is in the note, core Search, [[Omnisearch plugin for Obsidian|Omnisearch]], your operating system's file search and AI agents can all find it. And the callout is collapsible, so it doesn't clutter your notes. You pick the OCR engine: - **[[Tesseract]]** (default): free and fully local, but English only, images and PDFs only, and often less accurate - **Mistral OCR**: an [[Mistral AI]] cloud model that handles complex documents and handwriting, with an API key - **Any OpenAI-compatible API**, or a **custom command** of your own You can extract text from the active note, a folder or the whole vault, auto-extract new attachments, and reuse the text layer of a PDF instead of running OCR. On mobile, only the active note can be processed. Mind privacy: with a cloud engine, your images leave your machine. If you'd rather get a separate `.md` file next to each image, transcribed by a local vision model through [[Ollama]], take a look at my [[Transcriber plugin for Obsidian|Transcriber]] plugin. It's also a good fit for handwritten notes. ## References - Source code: https://github.com/jritzi/ocr-extractor - Community plugin page: https://obsidian.md/plugins?id=ocr-extractor ## Related - [[The Complete Guide to Obsidian Search - Find Any Note When You Need It (Article)]] - [[The Best Obsidian Plugins for Finding Notes (Article)]] - [[Transcriber plugin for Obsidian]] - [[AI Attachmate plugin for Obsidian]] - [[Text extractor plugin for Obsidian]] - [[Omnisearch plugin for Obsidian]] - [[Tesseract]] - [[Obsidian Callouts]] - [[Obsidian Community Plugins]]