How to Add Subtitles to a Video Automatically (2026)

To add subtitles to a video automatically, generate a timed transcript with speech recognition, review it against the audio, then export a subtitle file (SRT or VTT) or burn the text into the picture. YouTube Studio and OpenAI's open-source Whisper produce automatic subtitles at no charge; CapCut, Premiere Pro, DaVinci Resolve Studio and Loopdesk do it inside an editor.

This guide gives the exact clicks or commands for tools you may already have. "Automatic subtitles", "auto subtitles" and auto captions are the same speech-to-text step under different names. To choose a tool or compare prices, see our auto caption generator roundup; for styled social captions, see how to add auto captions to any video.

Facts checked November 13, 2026 against YouTube Help, CapCut's help center, Adobe's Premiere help, the DaVinci Resolve 21 Reference Manual and OpenAI's Whisper repository. Menus, plan gates and prices change, and we did not run these tools, so confirm in your own app. Full disclosure: Loopdesk is our product and appears as one of six routes, with its limits stated.

What do you need before you start?

You need three things: audio clean enough for speech recognition, the language being spoken, and a delivery decision: subtitles burned into the picture, or a separate sidecar file such as SRT or VTT. Settle these first, because they decide which tool and export setting you use.

• Clean audio. YouTube says automatic captions may not generate when a video "has poor sound quality" or has "multiple speakers whose speech overlaps".
• The spoken language. Most tools ask for it or detect it; check that yours is supported.
• Burned-in or sidecar. Compare:

Our SRT vs VTT guide covers the formats; how to burn subtitles into a video covers burning in an existing file.

How do you add subtitles to a video automatically in each tool?

Every tool follows the same pattern: run speech recognition on the audio, correct the text, then export. The steps below come from each tool's own help page, manual or README, read on November 13, 2026. They are in no ranked order, and where a page is silent we say "Not stated".

YouTube Studio: edit and download auto generated subtitles

YouTube creates automatic captions for long-form videos and Shorts and publishes them when they are ready.

1. Upload the video. Sign in to YouTube Studio and select Subtitles from the left menu.
2. Click the video, then select Edit for the language in the "Subtitles" column.
3. For automatic captions, select Duplicate and edit, then Continue.
4. Fix the text in any line, adjust the timestamp boxes if the timing is off, then click Publish.
5. To get the file, click Edit next to the language, then Options and Download subtitles.

The help page does not name the download formats (Not stated). The YouTube Data API's captions.download method lists sbv, scc, srt, ttml and vtt, for videos you have permission to edit.

CapCut: auto subtitles, then Caption Export

CapCut's help center (January 30, 2026) documents Auto Captions, also called "Recognise Subtitles", on the web, desktop and mobile. On desktop:

1. Open your project.
2. In the left toolbar, click Captions, then Auto Captions.
3. Select the audio source and language, then click Generate.
4. Adjust timing, text or appearance in the timeline.
5. When you export, choose Caption Export instead of Video Export and pick SRT or TXT. Only Auto Captions export; typed text does not.

Does SRT export need a paid plan? CapCut's help center does not say (Not stated). A June 2025 CapCut article says you can download the SRT file "if you have the pro version of CapCut"; a November 2026 one says the option appears "only if shown" for your region and version. If it is missing, see our CapCut alternatives.

Adobe Premiere Pro: transcribe, create captions, export an SRT file

Adobe's current help pages call the app Premiere, and advise adding captions "when your edit is finished or close to finished."

1. Select Window > Text. In the Transcript tab, open Generate static transcript.
2. Set Language, Speaker labeling and Audio analysis, then select Transcribe.
3. Correct the transcript, then select Window > Workspaces > Captions and Graphics.
4. Select Captions > Create captions from transcript, pick a preset (Adobe says "Subtitle default" suits most cases) and one or two lines, then select Create captions.
5. Select Export in the header bar. In the Captions section, set Export Options to Create Sidecar File and the file format to SubRip Subtitle Format (.srt).

For the file alone, the Text panel's Captions tab has Export to SRT file in its ellipsis menu.

DaVinci Resolve Studio: Create Subtitles from Audio

The DaVinci Resolve 21 Reference Manual marks this function "(Studio Version Only)", so the free DaVinci Resolve does not include it. Blackmagic lists DaVinci Resolve Studio at US$295 ($236 on sale on November 13, 2026); DaVinci Resolve Free vs Studio covers the other differences.

1. On the Edit page, open the timeline. Mark an In-Out range, or leave it blank for the whole timeline.
2. Select Timeline > AI Tools > Create Subtitles from Audio.
3. Set Language (Auto detects it), Caption Preset, Max Characters per Line and Lines, then click Create.
4. Choose File > Export Subtitle, or right-click the subtitle track header and choose Export Subtitle, and save as .srt or .vtt.

OpenAI Whisper: generate subtitles from audio

Whisper is OpenAI's open-source speech recognition model; its README says the code and model weights "are released under the MIT License." It runs from the command line on your own computer and needs Python and FFmpeg.

pip install -U openai-whisper
whisper interview.mp3 --model turbo --language English --output_format srt --output_dir subtitles

• --output_format accepts txt, vtt, srt, tsv, json, jsonl or all (the default, which writes every format).
• --output_dir sets the folder, and --language is optional: without it, Whisper detects the language.
• --max_line_width and --max_line_count limit line length and require --word_timestamps True.

Option names are from the README and whisper/transcribe.py. The README warns that "Whisper's performance varies widely depending on the language", and the command line has no option for speaker labels.

Loopdesk: ask Aura to transcribe

Disclosure: Loopdesk is our product. It is an agentic AI video editor that runs in the browser, and Aura is its AI agent.

1. Import the video (MP4, MOV, MKV, AVI and more, or from Google Drive or Dropbox).
2. Ask Aura in plain English to transcribe it, or use the Transcribe Quick Action.
3. Review the captions and correct names, numbers and jargon. Caption editing is unlimited.
4. Burn the captions into the video, or export an SRT or VTT sidecar file.

Loopdesk captions in 108 languages, including mixed-language videos and right-to-left scripts. The limits: it is paid only, with no free plan or free trial (Creator Lite is $19/month, or $15/month billed annually; see pricing), and browser only, so it needs an internet connection.

How do the automatic subtitle generators compare?

They differ in cost, where they run and what they export. YouTube Studio and Whisper cost nothing, CapCut is free to start, and the other three are paid. All six document a route to an SRT file, though YouTube names it only in its API reference.

"Listed" is our count of the vendor's list; "Not stated" means its official pages do not say.

How do you check automatic subtitles before publishing?

Play the video with the subtitles on and read every line against the audio. The machine's output is a draft: W3C says automatically generated captions "do not meet user needs or accessibility requirements, unless they are confirmed to be fully accurate." Check names, numbers, jargon, timing and line length.

• Names and brands. List the people, products and places, then search the track for each.
• Numbers. Compare prices, dates and units with the audio and on-screen graphics.
• Jargon and small words. Fix acronyms, and watch for a dropped "not", which W3C notes can make captions contradict the audio.
• Timing. Each line should appear as the speaker starts. Check again after any re-cut.
• Line length. Netflix's English style guide allows 42 characters per line and a "Maximum two lines."

For accessibility rules, see are auto captions ADA and WCAG compliant? This is not legal advice.

How do you convert or fix a subtitle file?

Use a converter when a tool hands you the wrong format or every line is early or late by the same amount. Our free SRT converter turns SRT into VTT, VTT into SRT and either into plain text, and can shift a whole file's timing. Typos need only a text editor.

YouTube accepts .srt and .vtt uploads and requires plain UTF-8 for SRT. Keep the reviewed file: fixing a sidecar file takes seconds, while removing hardcoded subtitles means cropping, blurring or inpainting.

What if the speech mixes languages or needs translating?

Mixed-language speech is a known weak spot. YouTube lists "multiple languages at the same time" among the reasons automatic captions may not generate, and Whisper's --language option takes a single language. Loopdesk's captions cover mixed-language videos; our multilingual captions guide explains how.

Translation is a separate job: Whisper's --task translate only translates speech into English, and Loopdesk does not translate. See how to translate a video with subtitles or AI dubbing, and captions vs subtitles vs SDH for the terms.

Where does Loopdesk fit?

Loopdesk (full disclosure: our product) does this job: Aura transcribes, you review, and you burn the captions in or export SRT or VTT. It fits best when subtitles are one step in a larger edit, since the same agent removes silences, reframes to 9:16 and works across 50+ videos.

It is paid only (metered AI credits, no free plan or free trial) and browser only, with no documented import for an existing SRT or VTT file. If the video is going to YouTube anyway, YouTube Studio is the better choice and costs nothing. To transcribe on your own computer, use Whisper.

Frequently Asked Questions

How do I automatically add subtitles to a video for free?

Two routes cost nothing: edit YouTube's automatic captions in YouTube Studio, or run OpenAI's open-source Whisper on your own computer to write an SRT file. CapCut is free to start, but SRT export may need a paid plan.

Can I generate subtitles from an audio file?

Yes. OpenAI's Whisper transcribes audio files such as MP3, WAV and FLAC, and its --output_format srt option writes a subtitle file. In Premiere Pro or DaVinci Resolve Studio, put the audio on a timeline first.

How accurate are automatic subtitles?

Good enough to start from, not to publish unchecked. YouTube says automatic captions can misrepresent speech because of mispronunciations, accents, dialects or background noise, and W3C says they usually need significant editing.

What is the difference between subtitles and captions?

W3C notes that the two words are used for the same thing in different regions. In its own usage, captions are in the language of the audio and include non-speech sounds, while subtitles translate the speech for people who can hear it.

Can I download YouTube's automatic subtitles as an SRT file?

You can download caption tracks for your own videos: in YouTube Studio open Subtitles, click Edit next to the language, then Options and Download subtitles. YouTube's help page does not name the formats; its Data API lists srt, vtt, sbv, scc and ttml.

Should I burn subtitles in or upload a subtitle file?

Upload a subtitle file wherever the platform accepts one, such as YouTube, so viewers can switch subtitles off and you can fix a typo without exporting again. Burn subtitles in where you cannot upload a file or the styling must travel with the video.

---

Need subtitles as part of a bigger edit? Open Loopdesk, ask Aura to transcribe the video, then export SRT or VTT. Plans start at $19/month ($15/month billed annually); there is no free plan or free trial.