Text-Grab

Text Grab

Copy any text you can see.

## Overview Use Text Grab on Windows to capture text with OCR, clean it up quickly, and move it into the next step of your workflow. When text gets trapped inside images, videos, PDFs, and parts of apps where you cannot select it, Text Grab helps you get it back out. You can take a screenshot or open a supported file, run it through the OCR engine, and send the result to the clipboard or straight into an editor. All OCR runs entirely on your device — no internet connection, no cloud service, and no per-use cost. Everything stays local and private. You can also do much more than copy text from images, because Text Grab gives you multiple modes for capture, post-grab cleanup, spreadsheet-style editing, and fast text reuse. You can use it all day without friction. Launch it quickly from the taskbar, open specific modes from the command line, or enable the background process so global hotkeys work anywhere in Windows. The Full-Screen Grab mode is also the basis of the [PowerToys Text Extractor](https://learn.microsoft.com/en-us/windows/powertoys/text-extractor). ### Requirements - **Windows 10 or later** — required for all features using the Windows OCR API. - **Windows 11 on a Copilot+ PC (Microsoft Store install)** — required for Windows AI features, which use the on-device Neural Processing Unit for higher-accuracy recognition. ## How to Install ### Official - [Microsoft Store](https://www.microsoft.com/en-us/p/text-grab/9mznkqj7sl0b?cid=TextGrabGitHub) - [GitHub Releases](https://github.com/TheJoeFin/Text-Grab/releases/latest) ### Community - [scoop](https://scoop.sh/) — `scoop install text-grab` - [choco](https://community.chocolatey.org) - `choco install text-grab` ## How to Build Build and test Text Grab on Windows. Get the code: - Install Git: https://git-scm.com/download/win - `winget install git.git` - `git clone https://github.com/TheJoeFin/Text-Grab.git` ### With Visual Studio 2022 or Visual Studio 2026 - Install Visual Studio (the free Community edition is sufficient). - Install the "Universal Windows Platform development" workload. - Install the ".NET desktop development" workload. - Install the ".NET cross-platform development" workload. - Install Windows 10 SDK `10.0.22621.0` - Open `Text-Grab.sln` in Visual Studio. - Set `Text-Grab-Package` as the startup project. - Set the CPU target to `x64` or `ARM64`. - Press `F5` or choose **Local Machine**. ### With the .NET SDK or Visual Studio Code - Install the .NET 10 SDK: https://dotnet.microsoft.com/download/dotnet/10.0 - This repository pins SDK `10.0.100` in `global.json`. - Optional for debugging: install Visual Studio Code https://code.visualstudio.com/ and the C# extension / C# Dev Kit. - Open the `Text-Grab` folder in VS Code. - Restore dependencies with `dotnet restore Text-Grab.sln` - Build with `dotnet build Text-Grab\Text-Grab.csproj` - Run tests with `dotnet test Tests\Tests.csproj` - In VS Code, press `F5` to launch with the included debug configuration. ### UI automation The Windows UI automation inventory, safe local release sign-off command, runner requirements, and opt-in system/package lanes are documented in [UiTests/README.md](UiTests/README.md). ## Choose from Four Modes ### 1. Full-Screen Mode (basis of [Text Extractor](https://learn.microsoft.com/en-us/windows/powertoys/text-extractor)) Use Full-Screen Mode when you want to select any region of the screen and copy the recognized text straight to your clipboard. You can also click once to try to copy a single word. That works because the Windows OCR API draws a bounding box around each recognized word. If you click or select an area with no text, the Text Grab window stays active so you can try again. To exit, press Escape, right-click and choose Cancel, or press Alt+F4. ### 2. Grab Frame Mode Use Grab Frame when you want a movable OCR window you can keep over part of your screen. Position the frame over the text you want, then grab text by searching for it, clicking a word border, or clicking the Grab button. Grab Frame uses the same OCR engine as Full-Screen Mode, so you get the same strengths and tradeoffs. OCR is not perfect, but you can often improve accuracy by adjusting the size and position of the frame. ### 3. Edit Text Window Use the Edit Text Window to turn OCR results into clean, usable content. You can work in plain text, spreadsheet-style, or markdown mode depending on what you need. Grab text with Full-Screen Mode or Grab Frame, then keep refining it without leaving Text Grab. Inside the Edit Text Window, you get tools that help you quickly fix, extract, and structure text. **Clean OCR output** - Make text into a single line - Toggle between UPPERCASE, lowercase, and Titlecase - Trim spaces and empty lines - Remove duplicate lines - Replace reserved characters (like spaces, /, %, etc.) - Extract text based on patterns like phone numbers, emails, or customer patterns **Search and extract** - Find and replace - Extract regular expressions - Launch URLs **Structure data** - Convert stacked data to table format - Continue working in Spreadsheet mode for row-and-column cleanup - Transpose captured table data when OCR gives the right data in the wrong orientation **Workflow helpers** - List files and folders in chosen directory - Watch clipboard for changes - Copy text from every image in a folder - Save or copy cleaned results into the next app in your workflow If you capture tabular data, use Spreadsheet mode to keep cleaning it after OCR. You can straighten columns, copy selected rows, copy the current column, and save the final result as delimited data for use elsewhere. The Edit Text Window also includes a **Calc Pane** for quick calculations alongside your text. It evaluates expressions in real time line by line, so you can stay in the same window while checking totals, conversions, or quick calculations from OCR results. The Calc Pane can help with: - Basic math like `+`, `-`, `*`, `/`, `%`, and exponents - Functions like `sin`, `cos`, `tan`, `sqrt`, `abs`, and `log` - Variables and constants like `x = 10`, `pi`, `e`, and `tau` - Unit conversions like `5 miles to km` or `100 fahrenheit to celsius` - Date math like `today + 5 days` or comparing dates in weeks - Aggregate values such as sum, average, median, count, min, max, and product Use the Calc Pane when OCR pulls numbers, quantities, dates, or measurements out of an image and you want quick answers without moving the data into another calculator first. It can also summarize numeric values from spreadsheet selections, copy all calculation results, and hide or show errors depending on how much detail you want while cleaning data. ### 4. Quick Simple Lookup Quick Simple Lookup is not about OCR. Use it to retrieve frequently used text like URLs, emails, part numbers, and more. Think of it as your long-term memory: a custom dictionary you can edit and recall instantly whenever you need it. The workflow is fast: 1. Press the hotkey (default is Win + Shift + Q) 2. Begin typing to filter the lookup to the item you want 3. When what you want is the first result, press Enter 4. Paste the value into the application you are using ### Bonus Mode: Browser Extension Text Grab isn't limited to screenshots. The [Text Grab Extension](https://github.com/TheJoeFin/Text-Grab-Extension) for Chrome and Edge lets you grab raw, selectable text straight from a web page and send it into Text Grab — no OCR needed since the text is already there. It's a fast way to pull text out of pages that block copy/paste or have awkward formatting, and get it into Text Grab for cleanup, spreadsheet mode, or your usual post-grab workflow. - [Chrome Web Store](https://chromewebstore.google.com/detail/text-grab-extension/edkbakekpnnghfmhojmckdagdjeldjmi) - [Edge Add-ons](https://microsoftedge.microsoft.com/addons/detail/text-grab-extension/mhhheckbddpoehjgcapihkmbjbgcgcfj) ### From OCR to usable text Text Grab keeps working after OCR. After each grab, you can run one or more post-grab actions to clean or route the text before it lands in the next tool. That helps when you repeatedly grab invoice line items, SKU tables, IDs, or lists from screenshots and want them in a cleaner, more usable shape right away. Examples of post-grab actions and workflow helpers include: - Trim each line - Remove duplicate lines - Fix GUID formatting - Try to insert text into the active app - Open a web search - Save reusable combinations of actions as grab templates One example workflow: capture a pricing table with Full-Screen Grab or Grab Frame, run a post-grab action like **Trim each line** to remove extra whitespace, then send the result into the **Edit Text Window** in **Spreadsheet** mode. From there you can correct OCR mistakes, move data into the right columns, transpose the table if needed, and then copy or save the cleaned result as data ready for Excel, Google Sheets, an ERP, or another step in your workflow. ## How Text Grab Captures Text Text Grab supports several capture methods so you can choose the one that best fits your hardware, workflow, and source material. All methods run entirely on your device — nothing is sent to a server, and there are no per-use costs. ### WinAI (Windows AI API) — Copilot+ PCs The newest and most accurate option. The Windows AI API is available on Copilot+ PCs with Text Grab installed through the Microsoft Store, and runs on the dedicated Neural Processing Unit (NPU). Because inference happens on dedicated silicon, it is fast and power-efficient. It produces higher-quality results than traditional OCR, especially on handwriting, stylized fonts, and complex layouts. WinAI models support a wide range of languages automatically without requiring additional language packs. If your device supports it, Text Grab will offer this as an option automatically. ### WinRT OCR (Windows OCR API) — Windows 10 and later The default capture method for most users. The Windows Runtime OCR API has shipped with Windows since Windows 10 and runs entirely on your device. It is fast, reliable, and produces excellent results for printed text in screenshots, documents, and images. Recognized languages depend on the language packs installed in Windows — add languages through Windows Settings to enable recognition in those languages. ### Tesseract An open-source OCR engine that Text Grab can use as an alternative to the WinRT API. Tesseract has been around for decades, is highly configurable, and supports a very large range of languages through downloadable language data files available in the [tessdata repository](https://github.com/tesseract-ocr/tessdata). It can be a good choice when you need more control over recognition parameters or when working with image types where you want to compare output between engines. ### Direct Text (UI Automation) Sometimes OCR is not needed at all. When text is displayed in a native UI element — a text box, label, list, or document rendered by the operating system — Text Grab can read it directly using Windows UI Automation without running any OCR. This approach is faster, perfectly accurate, and works regardless of display resolution or font. Use it when the text is selectable in theory but not easily accessible through normal copy and paste. ## Command Line Interface Use these arguments with `Text-Grab.exe`: - `Fullscreen` launches into Fullscreen Grab mode - `GrabFrame` launches a new Grab Frame - `EditText` launches a new Edit Text Window - `QuickLookup` launches Quick Simple Lookup - `Settings` opens Text Grab settings - `--grabframe "file path"` opens a supported image or PDF directly in Grab Frame - `--windowless