Video and caption editor for phones

By HYM Apps · Updated 8 November 2026

Guide: automatic captions with the Gemini API

  1. Open Google AI Studio with your Google account. The “Open Google AI Studio” button in the app also goes there.
  2. Tap Create API key, choose or create a project, and copy the key.
  3. In the app: open a project → Text → Auto captions, paste the key and tap Save key.
  4. Choose the language, tick the consent box and tap Make captions. While it runs you can switch apps or turn off the screen; if notifications are allowed, one shows the progress and has a Cancel button.

A long video is transcribed in one transcription request (audio longer than 3 minutes is uploaded in parts of about a minute each), and this can take several minutes. Half an hour of audio is an upload of about 58 MB, so Wi-Fi is recommended. If something fails, tick the consent box again and tap “Try transcription again” (audio without a saved result is sent again and may be charged again). If Google already answered, the answer is kept and “Check saved answer again” re-checks it without uploading; if this version of the app already could not use it, “Delete answer and transcribe again” deletes it and sends the audio again (it may be charged again). A 503 error from Google (service overloaded) usually clears after a few minutes. Before adding, the app flags long, overlapping or doubtful captions for review (if the answer for part of the audio cannot be used, the captions for the rest are kept); AI can make mistakes, so check the captions in the editor. In Caption display you can show the full subtitle, short phrases of up to 3 words or one word at a time, and hide commas and full stops. Under the Gemini API Additional Terms of Service, the service is for professional or business use, only for users aged 18 or over and only in supported regions; the app cannot verify your age, region or billing status. The Gemini API has an unpaid tier within limits set by Google, but in the EEA, Switzerland and the UK only the paid tier may be used, which means a project with active billing; if billing is on, use is charged to your Google project. On the unpaid tier, Google may use content to improve its products and humans may review it. Keep your key private; you can remove it in the app (Remove key) or revoke it in Google AI Studio at any time.

Privacy policy

In short: editing, export, projects and screen recordings stay on your phone. There is no app sign-up and we have no server. We (the developer) do not sell your personal data, and we receive it only when you choose to email us, for example a crash report or a support question. To have such an email deleted, write to webhym@gmail.com. The app shows ads from Google AdMob when it opens (see Ads below) and contains no analytics SDK. Apart from ads, the app itself sends data off the phone only when you use transcription with the Gemini API, as described here. When you choose to share, save a backup to a place such as Google Drive or translate text, the file or text goes to the app or service you picked, under its own policy. The same applies to a rating you submit in the rating window of the app store you installed from: it goes to that store, under its own policy.

Contact: webhym@gmail.com. When anything changes, we will update this page and its date.