Contexxt

Contexxt - In-Browser Media Workbench, Inspector & Offline OCR Inspect web media on any page, or open the local workbench to convert files and run OCR in your browser without uploading data. Contexxt provides an in-page overlay to inspect images, video, and audio alongside computed CSS styles. For file conversions and media edits, the workbench runs entirely in your browser. FEATURES 1. In-Page Asset & Style Inspector ● Right-click any element and choose "Inspect with Contexxt". ● Hold Ctrl (or Cmd) and click an asset to open its source in a new tab. ● Check natural and rendered dimensions, file formats, and estimated download sizes from HTTP headers. ● Inspect responsive media such as <picture> elements, srcset candidates, CSS background images, and video posters. ● View computed typography and styles, including font family, size, weights, CSS classes, and color swatches with hex copying. ● Download assets, open them in new tabs, or send them to workbench tools. 2. Image Conversion & Editing ● Convert between WebP, PNG, JPEG, and AVIF. ● Compress PNG files losslessly to reduce file size without losing quality. ● Rasterize SVGs at custom pixel dimensions or preset scales (1x, 2x, 4x). ● Resize with presets, crop, rotate, flip, and apply matte background fills. ● Compare before and after file sizes with compression savings percentages. 3. Audio & Video Editing ● View container and codec metadata for H.264, VP8/VP9, AV1, AAC, MP3, Opus, and FLAC. ● Transcode video to MP4 or WebM with custom resolutions and crop settings. ● Trim media using interactive range sliders. ● Export audio to MP3, AAC, FLAC, or WAV, or strip audio tracks from video files. ● Resample audio (Auto, 44.1 kHz, 48 kHz), mix down to mono, and set custom bitrates. ● Save video frames at any timestamp as PNG or JPEG images. ● Extract embedded cover art from audio files and pass it to image tools or OCR. 4. Offline Text Extraction (OCR) ● Extract text offline with built-in English recognition, requiring no network requests or cloud services. ● Recognize text from full images, cropped areas, or captured video frames. ● Choose layout analysis modes: automatic, single block, single line, or digits only. ● Preprocess low-contrast text with threshold and binarization filters. ● Copy text to the clipboard or export .txt and .tsv files with confidence scores. 5. Batch Processing & Chained Tools ● Drop multiple images into the workbench to convert and resize them in a single queue. ● The processing queue handles files independently so one corrupted file does not halt the batch. ● Download processed files individually or as a complete set. ● Pipe output from one tool directly into another with "Use as new input", such as passing a captured video frame directly into OCR. PRIVACY ● All media operations run locally on your device, so your files never leave your computer. ● Contexxt uses no remote conversion servers, analytics, telemetry, or user accounts. ● Permissions activate only when you trigger an action. ● Free and open source under GPL-3.0.