A new Birmingham, UK app transcribes your local video and audio with OpenAI's Whisper running on-device — private subtitles, a hover dictionary, and one-click Anki export, in 99+ languages.
2026-08-28 · by Moe Ameen
SubSmith, a desktop app from a small Birmingham, UK team, has surfaced in 2026 — listed on startup directories including BetaList and EU-Startups — with a specific pitch: learn a language from the content you already have. You drop a local video or audio file into the app and it transcribes the speech using OpenAI's Whisper model running on your own machine, then builds a study workflow around the result.
The app's defining choice is that everything runs locally. Because Whisper runs on-device, files never upload, there are no upload-size limits, and there are no per-minute cloud transcription fees. It handles common formats including MP4, MKV, and AVI video and MP3 and WAV audio, supports 99+ languages, and produces a timestamped, editable transcript. Around that transcript it adds a hover dictionary for instant word definitions, one-click export of a word plus its definition and the matching audio clip to Anki flashcards, and subtitle export as SRT or plain text.
SubSmith is positioned explicitly as a desktop alternative to browser tools like Language Reactor — it works on any local file rather than depending on a streaming platform or an extension — and it also markets to creators who simply want accurate, private subtitles for a video on their drive. Pricing centers on a low annual plan and a one-time lifetime license, with a 7-day trial that needs no credit card.
It's a new product, so treat specifics — operating-system support, format depth, the exact language count, and pricing — as a launch-window snapshot. SubSmith's own site (subsmith.app) is the source of truth as details settle.
There are two ways for a creator to act on this. The first is to treat it as a live content moment: the move toward local, on-device AI is a topic your audience cares about, so drop "a new app runs Whisper locally to subtitle your own videos privately" into Kompozy's Quick Ingest and fan one take into a Text Post, an X thread, a LinkedIn carousel, a captioned Short, and a blog explainer in your own voice — then schedule and publish the set across the eight social platforms plus blog and email while the angle is fresh.
The second is the more durable read: SubSmith is a clean example of the emerging two-stage creator stack — generate or prepare locally, distribute at scale in the cloud. SubSmith privately produces an accurate target-language subtitle track; Kompozy takes that finished video and does everything after the subtitle, cutting captioned vertical shorts, building vocabulary carousels and quote graphics, drafting a blog and newsletter, and publishing the batch across nine platforms on one governed brand voice. The local tool keeps your files private and your transcription costs at zero; the engine turns one subtitled lesson into a week of published content. Neither replaces the other — together they're a full pipeline.
SubSmith is a new desktop app from a Birmingham, UK team that turns your own local video and audio into language-learning material. It transcribes files with OpenAI's Whisper running on your device, adds a hover dictionary and Anki export, and supports 99+ languages.
Locally. Whisper runs on your own machine, so files never upload — there are no upload-size limits and no per-minute cloud transcription fees. Privacy and offline processing are its main selling points.
No. SubSmith produces subtitles, transcripts, and study cards and stops at an editable SRT. To clip that video into shorts, repurpose it into carousels or blogs, and publish across platforms, you'd pair it with a content engine like Kompozy.