Subtitles, on any video.

1. 1

   Open the side panel

   Pin the OmniSubs icon to your browser toolbar and click it on any page that has an HTML5 video element. The side panel lists every video it found in the active tab — including videos sitting behind tabs, carousels, and lazy-loaded sections.

2. 2

   Pick a target language and tone

   Choose any of 73 target languages, then pick a tone: natural, formal, casual, technical, or literal. Re-translating is free for 24 hours — change tone or language without re-running transcription.

3. 3

   Click start

   The extension captures the video’s audio in 10-minute chunks and uploads each chunk to omnisubs.app/api/transcribe. The original video file is never uploaded.

4. 4

   We transcribe and translate

   OpenAI Whisper transcribes each chunk; Google Gemini translates the resulting cues into your chosen language while preserving timestamps. Hallucinations and Whisper’s common boilerplate phrases are filtered out before subtitles ever leave the server.

5. 5

   Subtitles appear over the player

   The finished VTT file is rendered back as a transparent overlay anchored to the video element. Refreshing the page or navigating within the same site re-applies the cached subtitles automatically.

The extension can caption any HTML5 video element whose stream isn’t protected by Encrypted Media Extensions. We refuse DRM-protected players up front rather than letting users discover it the hard way.

• Only selected audio, subtitle text, subtitle files, and job metadata leave your browser. Video files are never uploaded in full.
• No ads or tracking in the extension. No third-party analytics, ad trackers, telemetry pings, or remote code. The full policy names every processor and host involved.
• User-initiated work only. Page/video data goes to OmniSubs only when you start a job, search/download subtitles, or submit a report. The side panel never reads passwords, form fields, or page text.
• DRM streams are refused, not bypassed. Encrypted Media Extension videos (Netflix, Disney+, Prime Video, etc.) are detected up front; the extension surfaces a clear refusal and stops.
• Auth tokens stay in session storage. Bearer tokens live in chrome.storage.session and are cleared when the browser closes. They are never shared with websites you caption.
• You own your transcripts. Generated subtitles are stored under your OmniSubs account and can be deleted at any time from your generations history.

Chrome shows these at install. Here is what every permission is used for, line by line.

storage

Saves your language and tone preferences, applied subtitles per tab, and the session-only auth token.

activeTab + scripting

Reads video elements in the tab you are actively using and injects the subtitle overlay onto that page. No cross-tab access.

alarms

Refreshes the short-lived auth token every 30 minutes so long jobs don’t stall partway through.

webNavigation

Clears the per-tab video registry when you navigate so stale thumbnails from the previous page don’t show up in the side panel.

webRequest

Observes (does not modify or block) media-asset network requests so we can recognize HLS / DASH manifests and progressive MP4s sitting alongside video elements.

sidePanel

Hosts the extension UI as a side panel that stays open while you interact with the player.

Host permission <all_urls>

Detects video elements on whatever site you choose to caption. Each tab is handled in isolation; there is no cross-site correlation, and page/video data is sent to OmniSubs only when you start a job, search/download subtitles, or submit a report.