Pick a video in your Media Library, press one button, and get a subtitle file\nattached to it. Visitors then see a subtitles button on the video player.<\/p>\n\n
The speech-to-text runs in the browser<\/strong>, on the machine of whoever pressed\nthe button. Not on your server, and not on anyone's API. That single decision\nis what this plugin is, and everything below follows from it.<\/p>\n\n There are two ways to turn speech into text: send the audio somewhere and get\nthe words back, or run the model where the audio already is. This plugin does\nthe second.<\/p>\n\n For a marketing clip either way is fine. For patient education, legal\nrecordings, internal training, HR material or paid membership content, sending\nthe file to a third party is often the difference between allowed<\/em> and not\nallowed<\/em>.<\/p>\n\n The browser reads the file, works on it, and writes the result back. The only\nthing that ever downloads is the speech model itself, once.<\/p>\n\n Or download Captions for prerecorded video are a WCAG 2.1 requirement (Success Criterion\n1.2.2). If your site has an accessibility obligation, this is one of the items\non the list.<\/p>\n\n Most of this would normally be left out of a plugin page. It is here because\nthe wrong expectation wastes your time and mine.<\/p>\n\n The two files in To rebuild from that source:<\/p>\n\n which regenerates Bundled into those two files by the build, both unmodified:<\/p>\n\n Two things are downloaded into the browser the first time you generate\nsubtitles, and then kept by the browser's cache. Nothing is uploaded, no\naccount is needed, and no key is required.<\/p>\n\n Your video is never sent to either of them, or to anywhere else. It is read\nfrom your own Media Library by your own browser and worked on there.<\/p>\n\n Terms and privacy: jsDelivr \u2014 https:\/\/www.jsdelivr.com\/terms and\nhttps:\/\/www.jsdelivr.com\/terms\/privacy-policy-jsdelivr-net \u00b7 Hugging Face \u2014\nhttps:\/\/huggingface.co\/terms-of-service and https:\/\/huggingface.co\/privacy<\/p>\n\n If your site cannot reach an outside server \u2014 some government, medical and\ninternal deployments cannot, and jsDelivr is blocked in some countries \u2014 put\nthe files on your own host and point the plugin at them:<\/p>\n\n The files to copy are the contents of Be honest with yourself about this before installing: if the first run cannot\nreach these two servers, the plugin cannot generate subtitles. After that\nfirst run it works offline.<\/p>\n\n\n Defaults live under Settings > Bharti Subtitles<\/strong>.<\/p>\n\n\n No. Nothing is sent to any API. The model runs inside your own browser.<\/p><\/dd>\n No. The browser reads the file straight from your own Media Library and works\non it locally. The only thing downloaded is the speech model itself, once.<\/p><\/dd>\n The speech model has to be downloaded the first time \u2014 about 50 MB for most\nlanguages, about 250 MB for Hindi, Marathi and Hinglish, which need a larger\none. After that the browser has it and there is no wait. Transcription speed\nthen depends on the machine, not on your hosting \u2014 a modern laptop does roughly\nreal time or better.<\/p><\/dd>\n Not in this version. Everything is deliberately in the browser, which is what\nkeeps the cost at zero and the video private. Bulk generation would mean a\nserver doing the work, and that means both of those go away.<\/p><\/dd>\n Check three things: the video has subtitles saved (Media > Bharti Subtitles will say\nso), Settings > Bharti Subtitles > \"Automatically show subtitles\"<\/strong> is on, and your\ntheme is not replacing the video player with its own. If it is, turn the\nsetting off and add the track yourself.<\/p><\/dd>\nNothing leaves the computer<\/h4>\n\n
No account, no key, no bill<\/h4>\n\n
\n
How it works<\/h4>\n\n
\n
.vtt<\/code> is written beside the video and\nattached to it<\/li>\n<\/ol>\n\n.srt<\/code> \/ .vtt<\/code> instead and take it to YouTube, Vimeo or a video\neditor.<\/p>\n\nWhat you get<\/h4>\n\n
\n
[video]<\/code> shortcode<\/li>\n.pot<\/code> file included<\/li>\n<\/ul>\n\nAccessibility<\/h4>\n\n
Honest limits<\/h4>\n\n
\n
.mov<\/code> will not decode.<\/li>\n<\/ul>\n\nWhat it needs<\/h4>\n\n
\n
.vtt<\/code> is saved beside the video<\/li>\n<\/ul>\n\nSource code and build<\/h3>\n\n
assets\/js\/<\/code> are built, not written by hand. Their complete\nsource ships inside this plugin so nothing needs to be taken on trust:<\/p>\n\n\n
src\/<\/code> \u2014 all of the TypeScript and React that becomes them<\/li>\npackage.json<\/code>, tsconfig.json<\/code>, vite.config.ts<\/code>, vite.worker.config.ts<\/code> \u2014\nthe build configuration<\/li>\n<\/ul>\n\nnpm install\nnpm run build\n<\/code><\/pre>\n\nassets\/js\/subtitles.js<\/code> and\n assets\/js\/transcribe-worker.js exactly as shipped.<\/p>\n\nThird-party code<\/h4>\n\n
\n
External services<\/h3>\n\n
The speech engine<\/h4>\n\n
\n
https:\/\/cdn.jsdelivr.net\/npm\/@huggingface\/transformers@3.8.1\/dist\/<\/code><\/li>\nThe speech model<\/h4>\n\n
\n
https:\/\/huggingface.co\/<\/code><\/li>\nServing them yourself<\/h4>\n\n
add_filter( 'bharti_subtitles_wasm_paths', function () {\n return 'https:\/\/example.com\/onnx\/'; \n} );\n<\/code><\/pre>\n\n@huggingface\/transformers@3.8.1\/dist\/<\/code>.\nThe model can be self-hosted the same way; see the Transformers.js\ndocumentation for env.remoteHost<\/code>.<\/p>\n\n\n
\/wp-content\/plugins\/<\/code>, or install the ZIP\nthrough Plugins > Add New > Upload Plugin<\/strong>.<\/li>\n\n
Do I need an OpenAI key or any other account?<\/h3><\/dt>\n
Is my video uploaded anywhere?<\/h3><\/dt>\n
Why is the first run slow?<\/h3><\/dt>\n
Can I run it on the server instead, or in bulk overnight?<\/h3><\/dt>\n
The subtitles do not show on the front end.<\/h3><\/dt>\n