Three subtitle tools. Pick the one you need.
Extract burned-in subtitles, translate subtitle files, or transcribe speech — each tool is built for one clear workflow.
macOS 13.0+ · Apple Silicon (M1/M2/M3/M4) only, Intel not supported
Windows · 64-bit desktop build
7-day free trial of all Pro features — no sign-up or login ever needed
Different videos create different subtitle problems. GeekLink gives you a local workflow for recovering text from the picture, transcribing speech, translating subtitle files, and exporting captions you can actually edit. Start with the format you have, then keep the timing and subtitle work together as your project grows.
Extract burned-in Japanese, Chinese, and Korean subtitles from anime, variety shows, and fan-subbed releases, translate them, and create editable subtitle files. OCR is useful when the video has visible captions but the original subtitle track or SRT is missing. Review names, lines, and timing before you share or reuse the subtitles.
Extract anime and fansub subtitles with OCRRecover subtitles when you cannot find the original SRT — from foreign films, dramas, TV episodes, and older media. GeekLink can work from hardcoded captions with OCR or from spoken audio with transcription, giving you a practical starting point for translation, archiving, and accessible subtitle versions.
Recover movie and TV subtitles as SRTTranscribe long-form training videos, lectures, coaching sessions, and online courses, translate them into another language, and burn the translated subtitles back into MP4. Process a module at a time or work through a larger course library while keeping the subtitles time-aligned and editable.
Transcribe training and course videosBatch-process videos through transcription, translation, subtitle editing, and export. Creators localizing a channel, short-form library, documentary collection, or recurring series can move from source video to translated or bilingual SRT, then produce a finished subtitled video when needed.
Translate subtitles for video localization"I couldn't find complete Chinese subtitles anywhere else. GeekLink extracted them in minutes — amazingly fast and surprisingly accurate."
"Overall I'm very satisfied with the performance. The translation speed and transcoding performance on an M4 Mac are excellent."
"I was looking for OCR subtitle software for Mac with an easy-to-use interface that was reasonably fast and accurate. I tried a few and found GeekLink the best, especially with Chinese subtitles."
"Recognition of Chinese subtitles is excellent — faster than voice processing."
A hardcoded subtitle is text rendered directly into a video's picture. It is also called a burned-in or open subtitle: it is part of the video image rather than a separate selectable subtitle track, so it cannot be turned off or edited until you extract it with OCR.
Yes. GeekLink OCR detects visible subtitle lines in a video and converts them into editable, time-synced SRT subtitles. It can work even when the video has no separate SRT file or usable audio track.
Yes. GeekLink can transcribe speech from a video with local speech recognition, even when the video has no existing subtitles. The video needs a usable spoken-audio track; after recognition you can edit the timing and text and export an SRT.
Yes. Import an existing SRT file, choose the target language, and export a translated or bilingual SRT. You can also translate subtitles created through OCR or speech recognition and burn the result back into video.
Speech recognition and OCR run locally after their models have been downloaded, so those workflows can run without an internet connection. AI translation needs internet because the selected translation model runs through a cloud service.
Yes. Import multiple videos and let GeekLink process a season, course library, or other collection through the subtitle workflow. The desktop app queues the videos so you can transcribe, translate, edit, and export them in batches rather than repeating the setup for every file.
Yes. GeekLink provides a Mac build for Apple Silicon Macs running macOS 13.0 or later and a 64-bit Windows desktop build. Intel Macs are not supported.