Extract Hardcoded Subtitles

Read subtitles burned into a video and save them as an editable subtitle file

Drop your video here

or click to browse

Video: MP4, MKV, WEBM, MOV, AVI, M4V

How to extract hardcoded subtitles

1

Choose your video and language

Open an MP4, MKV, WebM or MOV video and select the language of the visible subtitles.

2

Check the subtitle region

Confirm or adjust the suggested region so it covers the subtitles you want to read.

3

Read and review the subtitles

Start extraction, then use the To check list and video preview to correct text and timing.

4

Download the subtitle file

Choose SRT, VTT, ASS or TXT and save the result. SRT is selected by default.

Read subtitles from the video picture

Hardcoded subtitles are part of the video image. They cannot be selected or extracted as a subtitle track. This tool reads the visible text with OCR and turns it into editable, timed cues.

How burned-in subtitle extraction works

Choose the subtitle language and check the suggested subtitle region. The browser reads video frames, recognizes the text and groups it into timed cues. Review the result against the video before downloading a subtitle file.

Six subtitle languages

Chinese, Japanese, Korean, English, Indonesian and Turkish are supported. Choose the language of the subtitles on the picture, which may differ from the spoken audio. Vietnamese is not offered yet.

Your video stays on this device

The video is processed locally in your browser and is not uploaded. An internet connection is needed to load the page and download the OCR models on first use. The models are cached in this browser for later runs.

Browser and file limits

Speed depends on your device, browser and video. Firefox can be slower because GPU acceleration is not always available. Phones have less memory and may need shorter clips. For files over two hours, extract a shorter section first. Not every browser can decode every video codec.

Frequently Asked Questions

Is my video uploaded?

No. Video decoding and subtitle recognition run locally in the browser. The video stays on your device. The first run downloads the OCR models, which are then cached for later use.

Which subtitle languages can I extract?

Chinese, Japanese, Korean, English, Indonesian and Turkish are supported. Select the language of the text visible on the video. Vietnamese is not offered yet.

How accurate is hardcoded subtitle extraction?

Accuracy depends on the language, letter size, source quality and background. Use the To check list to find cues that need attention, then compare the text and timing with the video. Flags help prioritize review but do not guarantee that other cues are correct.

Which subtitle formats can I download?

Download SRT, VTT, ASS or TXT. SRT is the default. The result contains recognized text; original fonts, colors and animation are not preserved. TXT is useful when you only need the words.

Why can extraction take a while?

Reading subtitles from video frames takes more work than extracting an existing subtitle track. The first run also needs a one-time model download. Speed varies with the device and browser; GPU acceleration can help, while Firefox and phones may run more slowly.

Tips for readable results

Use the clearest source available. Small letters, heavy compression and moving backgrounds make OCR harder.

Keep the subtitle region close to the text. Logos, credits and signs can otherwise appear as unwanted cues.

Review the To check list and sample the rest of the video too. A cue without a flag can still contain a recognition or timing error.