GeekLink is an AI-powered subtitle editor built for Mac that handles the full subtitle workflow locally on your machine. It transcribes speech to text in 40+ languages using Apple Silicon acceleration, extracts hardcoded subtitles via OCR, translates with DeepSeek AI, and burns subtitles into video. You can process dozens of videos at once through its batch pipeline. Speech recognition and OCR run entirely offline, so your video files never leave your Mac. It suits content creators, translators, and video editors who need fast, private subtitle production.

• Offline Speech Recognition: Transcribes speech to text in 40+ languages directly on your Mac using Apple Silicon acceleration. No cloud upload required, keeping your content private.
• OCR Subtitle Extraction: Detects and extracts hardcoded subtitles from any video, including drama, anime, and screen recordings, using CoreML-accelerated OCR. Works even without an audio track.
• AI Translation via DeepSeek: Translates subtitles into 40+ languages with DeepSeek AI. Supports bilingual display with both source and target languages shown side by side.
• Batch Processing Pipeline: Import dozens of videos at once. The full pipeline (transcription, OCR, translation, export) runs automatically on every video in a single workflow.
• Inline Editing with Live Preview: Edit timing, text, and translations directly. Add, delete, merge, and split subtitle lines with a real-time video preview synced to your edits.
• Flexible Export Options: Export as soft subtitles (.srt), burned-in hard subtitles embedded in video, or plain text. Bilingual subtitles supported across all formats.

• Strong Privacy by Design: Speech recognition and OCR process entirely offline on your Mac. Video files are never uploaded to any server.
• All-in-One Workflow: Handles transcription, OCR extraction, translation, editing, and export in a single app, so you don't need to juggle multiple tools.
• Fast Batch Processing: Process dozens of videos through the full pipeline automatically, saving significant time on large projects.
• Bilingual Subtitle Support: Generates side-by-side source and target language subtitles in all export formats, useful for language learning content and international distribution.

• Apple Silicon Only:: Requires an M1, M2, M3, or M4 Mac. Intel Macs are not supported, which limits access for users with older hardware.
• No Windows Support Yet: Currently Mac-only. A Windows version is on the waitlist but has no confirmed release date.
• AI Translation Accuracy Varies: DeepSeek-powered translations may produce fluent but incorrect results for complex or nuanced phrases, requiring manual review and post-editing.

• Content Creator Subtitling: YouTubers and social media creators can auto-transcribe and translate their videos into multiple languages, then burn subtitles directly into the final export. This removes the need for separate transcription services and subtitle editing tools.
• Anime and Drama Fansubbing: Fans translating foreign-language shows can use OCR to extract existing hardcoded subtitles, then translate them into their target language. The bilingual display feature helps verify translation accuracy against the original.
• Corporate Video Localization: Companies producing training videos, product demos, or internal communications can batch process entire video libraries. The automated pipeline handles transcription and translation across 40+ languages in one pass.
• Podcast and Interview Transcription: Podcasters and journalists can transcribe recorded conversations offline, clean up filler words with the non-verbatim mode, and export clean text transcripts for show notes or articles.
• Language Learning Content: Educators creating bilingual learning materials can generate side-by-side subtitles showing both the spoken language and the learner's native language, then export them as SRT files or burned into video.
• Screen Recording Documentation: Developers and instructors recording tutorials can extract and edit subtitles from screen recordings, even those without narration, using the OCR engine to capture on-screen text.