SubSmith: Offline Desktop App Turns Personal Videos into Language-Learning Material
A developer on Hacker News introduces SubSmith, an offline-first desktop app that converts personal videos into language-learning tools with local transcription and Anki export.
According to the post, the developer built SubSmith after spending years learning Japanese and struggling with a repetitive workflow. They would repeatedly switch between a video player, subtitles or transcripts, a dictionary, screenshots, audio clips and Anki whenever they heard a sentence worth studying. SubSmith aims to bring that workflow into one place.
The app lets users drop a video or audio file into it, generate a transcript locally, and then work with the transcript alongside the media to look up words and sentences, replay individual lines, edit the transcript, save useful sentences with their original context and audio, and export them as Anki cards. The developer stressed that the tool works with any personal media and is not tied to a particular streaming service or library.
A key feature is that SubSmith is offline-first and transcription happens on the user's device rather than sending media to a transcription API. This means privacy and the ability to work with arbitrary files without internet dependence.
The developer said that with a solid core now in place, the priority is to understand where the workflow breaks down for other learners rather than adding features randomly. They invited feedback on several specific points: whether users would actually save sentences from their own media, which part of the process feels like too much work, whether having audio and context attached makes an Anki card more useful, whether the tool would be better as part of an existing video player or browser, whether installing a desktop app is a barrier, and whether requiring an account before starting the free trial makes people give up.
The current version does require an account to begin the trial, and the developer acknowledged this might be meaningful friction for potential users. They also noted that the app is free to try and particularly asked for input from people who already learn languages through their own videos, anime, films, podcasts or other media. The post appears on Hacker News, where the app has received one comment so far.