How to Add Hinglish Subtitles in CapCut

The auto-caption button will not do this well. Here is what does.

Blog · 10 August 2026 · 6 min read

CapCut has auto-captions built in, and for a video entirely in English they are decent. Point them at a sentence like "bhai yeh editing software literally free hai" and the result is usually unusable: half the Hindi is guessed at, English words are attached to the wrong ones, and the timing wanders.

This is not a CapCut flaw exactly. It is what happens when speech recognition tuned for one language meets a sentence that switches languages twice. The fix is to caption outside CapCut and bring the result in, which takes about two extra minutes and gives you captions that actually read.

Why the built-in captions struggle

Speech recognition narrows its guesses using a language model: having heard "bhai yeh editing", it predicts what tends to follow in that language. Code-switching breaks the prediction, because the next word comes from a different language entirely.

Single-language engines resolve this by forcing everything toward the language they were told to expect. Set it to Hindi and English words come out as odd phonetic Hindi. Set it to English and the Hindi turns to noise. Neither setting is right, because the sentence is genuinely both.

The workflow

Step 1: finish your edit first

Do your cuts, trims, and speed changes before captioning. If you caption first and trim afterwards, every timestamp after the trim is wrong. This one habit prevents most subtitle sync problems.

Step 2: export the edited video

Export from CapCut at your normal settings. This exported file is what you will caption, so its timing matches what you will eventually publish.

Step 3: generate the captions

Upload that file to a tool that handles code-switching in a single pass. DesiCaptions transcribes Hindi, English, and any mix together, and writes the result in Roman script so it reads the way your audience types. A five minute video takes around 60 to 90 seconds.

Step 4: fix names and brands

Check the transcript for proper nouns, brand names, and slang. These are what any AI gets wrong most often. Correcting them now takes seconds and saves editing text on a timeline later.

Step 5: export an SRT

Download the SRT rather than the burned-in video, since you are going to style the captions inside CapCut.

Step 6: import into CapCut

• Desktop: Text, then Local Subtitle, then Import File, and select your .srt. Every cue arrives as a separate text clip on the timeline, correctly timed.
• Mobile: SRT import is limited or missing depending on version. If your version lacks it, the practical route is to export a burned-in video from the caption tool and skip CapCut captioning entirely.

Step 7: style once, apply to all

Style a single caption clip the way you want it, then use CapCut Apply to All so the rest match. Doing this after import means the styling survives, because the text is already final.

If your CapCut version will not import SRT

Mobile CapCut has shipped several versions with no local subtitle import. Rather than fighting it, invert the order: have the caption tool burn the captions into the video, then bring that finished video into CapCut for any remaining edits like music, effects, or transitions.

You lose the ability to restyle captions inside CapCut, since they are pixels at that point. You gain captions that can never drift out of sync, which for short-form is usually the better trade.

When the import fails

A few specific failures come up repeatedly, and each has a quick check.

• The file will not appear in the picker. CapCut filters by extension. A file saved as subtitles.srt.txt is invisible to it. Turn on file extensions in your file manager and check the real name.
• It imports but every caption is empty. Almost always an encoding problem. Re-save the file as UTF-8 without a byte order mark.
• It imports as one enormous block. The blank lines between cues are missing, so CapCut reads the whole file as a single caption.
• Captions land at the wrong times. The SRT was made from a different cut than the one on your timeline. Re-export the video and caption that file.

If none of those apply, the fastest route past it is to skip the import: export a video with the captions already burned in and use CapCut only for music and effects.

Settings that work for vertical video

Once the text is in, these are reasonable starting numbers for a 1080 by 1920 Reel or Short.

Style one caption clip with these, then Apply to All. Doing it in that order means you set it once rather than forty times.

The vertical position matters more than people expect, because the bottom of the frame is covered by Instagram own interface. The exact margins are in the Reels safe area guide.

Getting the look right

A few things that matter more than the font choice:

• Keep lines short. Three to five words per caption for short-form. A full sentence on screen is not read, it is skipped.
• Stay inside the safe area. Instagram and TikTok overlay their own interface on the bottom quarter of the screen. Captions placed there get covered.
• Use a heavy weight with an outline or shadow. Captions have to survive being played over a bright background at small size.
• Do not centre captions over faces. Slightly below centre is the usual sweet spot for vertical video.