The latest articles on DEV Community by Hard Work (@fastlog). https://dev.to/fastlog https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4025360%2Fe2ed8dd0-77b5-4ff4-bec8-8a51ac4acdab.jpg https://dev.to/fastlog en Hard Work Sat, 11 Jul 2026 16:20:40 +0000 https://dev.to/fastlog/a-browser-first-workflow-for-multilingual-video-subtitles-31i1 https://dev.to/fastlog/a-browser-first-workflow-for-multilingual-video-subtitles-31i1 <h1> A browser-first workflow for multilingual video subtitles </h1> <p>Video subtitle translation looks simple until you need the result to be readable on screen. A useful workflow has to keep timestamps, allow review, preserve terminology, and produce files that editors can reuse.</p> <h2> 1. Start with a timestamped transcript </h2> <p>Treat transcription as structured data rather than a paragraph of text. Each segment needs a start time, end time, and editable source sentence. This makes it possible to fix names, numbers, acronyms, and sentence boundaries before translation.</p> <h2> 2. Translate in subtitle-sized units </h2> <p>A literal sentence-by-sentence translation often creates lines that are too long for the available reading time. Split long thoughts at natural pauses, keep terminology consistent, and avoid breaking a phrase where a viewer would naturally read it as a unit.</p> <h2> 3. Keep original, translated, and bilingual outputs separate </h2> <p>Different viewers need different outputs. Language learners may prefer original-plus-translation captions, while editors may need a clean translated SRT. Keeping the outputs separate also avoids permanently burning one choice into the video too early.</p> <h2> 4. Review timing before rendering </h2> <p>A subtitle should begin with the spoken thought and remain visible long enough to read. If a translation needs more space, split it into shorter blocks instead of squeezing it into the source segment’s duration.</p> <h2> 5. Render only after the text is approved </h2> <p>When the content is correct, export an SRT for editing workflows and optionally render subtitles into an MP4 for immediate sharing. A browser-first pipeline can keep the media preparation local while using speech recognition and translation services for the language work.</p> <p>I built <a href="https://videotranslator.org/video-subtitle-translator" rel="noopener noreferrer">Video Translator</a> around this sequence: transcribe speech, translate timed segments, review the text, then download original, translated, or bilingual SRT files—or render the selected captions into an MP4. The key design choice is keeping the subtitle text editable until the final export.</p> <p>The result is a more practical multilingual workflow: captions that remain readable, reusable, and reviewable instead of being a one-way automatic conversion.</p> ai