You've finished editing a video, and you're ready to post. Then you watch it with the sound off. The introduction makes no sense, the demonstration loses its explanation, and the automatic captions turn your business name into something you'd never say.

You can add subtitles to video free, but getting text onto the screen is only part of the job. The practical approach is to choose a free route, correct the words that matter most, and check that viewers can read them without missing the action.

Free Subtitle Routes Compared at a Glance

In 2024, 50% of Americans reported turning on subtitles most of the time, according to Kapwing's compilation of subtitle-viewing research.

That makes captions useful even when viewers can hear the audio. Someone watching your property tour on a train or following a cooking demonstration in a noisy kitchen shouldn't have to guess what you're saying.

Free subtitle routes compared

The main difference between free tools is what you get at the end. A transcript is a document containing words. A subtitle file contains words and the timing instructions needed to display them alongside a video.

Free-plan limits change. Before uploading a long recording, check the current processing allowance, export options, watermark policy and privacy settings. Test an export before spending time correcting an entire transcript.

Choose by the job in front of you

• A short vertical clip: Use TikTok's caption editor when you're publishing directly there. Choose a browser editor if you need a downloadable captioned video for several destinations.
• A long YouTube upload: Start with YouTube's automatic track. Correct it in Studio, then download the caption file if you need to reuse it.
• An existing interview or recording: Start with transcription, then move the cleaned text into a timing editor.
• A video with exact styling requirements: Use Aegisub when positioning, font treatment and individual caption timing matter more than speed.

Think of these as five practical methods, rather than nine unrelated products: platform automation, browser captioning, transcript-first editing, manual subtitle authoring and long-video repurposing.

Most tools either generate words or produce timed captions. Some do both, so you don't always need to move between applications.

How to Auto-Generate Captions on YouTube and TikTok

The fastest starting point is usually the platform where you'll publish. Automatic captioning needs detectable speech, though. Don't expect a silent demonstration or music-only clip to produce descriptions of what viewers need to hear.

Menu labels change between devices and account versions. Follow the function, such as subtitles, captions or language settings, rather than expecting every button to match a tutorial.

Generate a YouTube caption track

1. Open the video in YouTube Studio. Select it under Content so you're editing the correct upload.
2. Open Subtitles and set the spoken language. Choose English if that's what the speaker uses, rather than the language you eventually want to translate into.
3. Look for the automatic caption track. Allow processing to finish if the upload is still being analyzed.
4. Open the draft for editing. Check the introduction first to catch a wrong language setting, missing speech or consistently incorrect terminology.
5. Save the corrected track. Use its options menu to download a caption file if another platform needs it.

YouTube doesn't necessarily offer a button that instantly creates automatic captions on demand. If no track appears, confirm the language and audio, allow more processing time, or use the transcript-first method instead.

This visual summarizes the publishing sequence.

Generate captions in TikTok

1. Open the caption tool after editing your clip. Generate captions from the version you intend to publish.
2. Choose automatic captions. Confirm the spoken language if the app offers that setting.
3. Edit each incorrect block. Fix wording and adjust timing wherever the available controls allow.
4. Choose readable styling and placement. Keep the text clear of faces, demonstrations and interface buttons.
5. Preview the published post. Check the viewing experience, not just the editing canvas.

A caption layout that looks fine inside TikTok may need repositioning on Instagram. The guide to adding captions to Instagram Reels is useful when preparing a separate version.

For a local service demonstration, start with audio that clearly explains the work. Boocoo's resource on affordable videos for local trades can help you think through the underlying production before captioning it.

Publishing rule: A generated track is a draft. Saving it successfully doesn't mean someone has checked what it says.

Use the quality checks below before treating either platform's output as publishable.

Caption Timing and Formatting Rules That Hold Up

A caption such as [door slams] Mara: Who left the gate open? communicates something that dialogue alone misses. The sound explains why Mara reacts.

The World Wide Web Consortium's caption guidance distinguishes accessibility captions, which include relevant speech and non-speech audio, from translated subtitles, which generally carry dialogue in another language. A translated transcript isn't automatically a complete accessibility track.

Caption formatting targets

Use these as working settings, then adjust them to the destination player and actual footage. They're editing targets, not a universal compliance specification.

Don't shrink the font to rescue an overloaded caption. Split the sentence at a natural phrase boundary instead. A phone viewer shouldn't have to choose between reading tiny text and watching the demonstration.

The visual below provides starter settings to test in your editor, rather than a substitute for checking your finished video.

Make each caption easy to follow

• Break by meaning: Keep an article with its noun and a modifier with the word it describes.
• Identify unclear speakers: Put a name or consistent label at the beginning of the relevant turn.
• Describe meaningful sounds: Include a warning alarm or an off-screen crash when it affects understanding.
• Preview on a phone: Watch at normal speed without pausing to finish reading.

A useful test is whether you can follow the scene with the sound muted. If a speaker reacts to an unheard event and the captions give no explanation, the track is missing information.

How to Turn a Transcript Into a Timed Subtitle File

When automatic captions are unavailable or unusable, separate the work: get the words, clean the words, then time the words. Correcting a product name in a document is easier than fixing it across a collection of carefully timed cues.

Manual review isn't busywork. A study of YouTube auto-captioned instructional video found 525 phrase-level errors across 68 minutes, illustrating why generated text needs checking before publication.

Keep the video, transcript and caption project together while you work.

Build the subtitle file

1. Get a transcript. Upload an eligible recording to Otter's free tier, or use Google Docs voice typing to dictate your own narration.
2. Clean the text in a document. Correct names, product terms, numbers and sentence boundaries before working on timing.
3. Load the video into Subtitle Edit. Import the cleaned plain text using its text-import tools, then inspect how the editor has divided the cues.
4. Time each cue against the waveform. Set its start at the relevant speech and its end after the phrase finishes, splitting long passages at natural breaks.
5. Save as SRT and test playback. Reopen the file alongside the matching video to confirm that your player displays it correctly.

Google Docs voice typing doesn't supply subtitle timestamps. It's useful for drafting narration, but it won't turn an uploaded interview into a synchronized caption track by itself.

Subtitle Edit's waveform helps you locate speech, but a visible peak isn't proof of a particular word. Listen while moving cue boundaries, especially around breaths, pauses and overlapping voices.

For help deciding which transcription approach suits your recording, compare the options in this video-to-text converter guide. Choose based on input support, export access and privacy, not just how quickly the first draft appears.

Choose specialist tools only when needed

Aegisub suits captions that need detailed positioning, fonts and colour, while Amara is useful when collaborators or translators need to edit a shared project. You don't need either to produce a readable SRT file.

This accompanying video offers another way to follow the workflow.

Keep an untouched copy of the transcript. It gives you something to compare against when an edit accidentally removes a phrase.

For an ordinary talking-head video, automation can handle the initial transcription. For specialized training or sensitive instructions, plan for someone who understands the subject to review the finished file.

Burned-In or Sidecar Captions and How to Export Them

Should every viewer see the captions, or should viewers choose whether to display them? That decision determines whether you need burned-in text or a separate caption track.

Burned-in captions become part of the picture. A sidecar file travels alongside the video and supplies text to a compatible player.

Caption delivery methods

For a website player, WebVTT is designed for browser-based timed text through the HTML track element, as explained in the Federal Agencies Digital Guidelines Initiative's WebVTT guidance. SRT is a practical choice when a platform or editing application asks you to upload a subtitle file.

Export and attach the correct track

Export the format your destination accepts. Name it with the video title and a language code, such as boiler-maintenance.en.srt, so you can distinguish it from translated versions.

Open the platform's subtitle or caption menu, upload the file, and select the matching language. Preview the uploaded result because the platform may render line breaks or positioning differently from your editor.

For a burned-in Kapwing clip, upload the finished video, open its subtitle tool, generate the draft and correct the text before exporting with captions visible. Check the free export conditions before investing in detailed styling.

The choice between a subtitle track and designed on-screen text is also useful when comparing text-in-video software. Animated headlines can emphasize a message, but they shouldn't replace the complete spoken content when you need captions.

Keep a master: Save the editable caption file even when your final delivery is burned into the video.

Once text is rendered into the picture, changing a spelling mistake requires another render. Also recheck synchronization after trimming the footage or changing the video's playback timing.

Add a cleaned transcript or a useful summary to the description where appropriate. That gives viewers readable context without forcing them to hunt through the video for a service detail or instruction.

Quality Checks and Fixes for Common Caption Errors

A wrong address or missing not can cause more trouble than an awkward line break. Start with mistakes that could change what someone does, buys or believes.

YouTube's automatic-caption guidance warns that accents, dialects, mispronunciations and background noise can cause speech to be misrepresented, and recommends reviewing the generated captions. Clear-looking sentences still need checking because a plausible word can be completely wrong.

Review by consequence, not just sequence

Search the transcript for names, addresses, prices, technical terms and calls to action. Check each occurrence against the audio, including repeated mentions where the recognizer may have made different guesses.

Then inspect difficult passages: background music, interruptions, speakers talking over each other and quiet speech. Finish by checking readability on a phone, including the beginning, transitions and ending.

Risk-based review helps you prioritize; it doesn't certify unchecked passages. For high-stakes material or a complete accessibility deliverable, verify the full track against the recording.

Caption symptoms, causes and fixes

Say a property video captions an agent's statement as “parking is included” when the audio says it isn't. The grammar looks fine, so a spelling check won't flag it. Only comparison with the recording catches the changed meaning.

Music-led clips deserve a separate listening pass because a caption generator may confuse lyrics, narration and background voices. If that's your format, AIMVG's overview of Veed Io for short-form music content can help you assess a music-oriented editing workflow, but the final text still needs human checking.

Decide whether free is enough

A free route is workable when you can access the full output, edit it, export the required format and complete an appropriate review. It stops being enough when limits prevent you from checking the whole recording, protecting confidential material or delivering what the audience needs.

Don't solve a review problem by buying another generator and assuming its output will be correct. If a wrong word could change a decision, price or diagnosis, a human checks that line before publication.

Translate Only the Parts That Need Review

General commentary and high-stakes instructions need different translation checks. A fluent reviewer can prioritize titles and key messages in low-risk content, but legal, health, financial, religious, technical or location-specific claims need a full contextual review.

Choose languages from your actual audience needs rather than publishing every available option. Cisdem's comparison of automatic-caption accuracy describes differences across speech samples, a reminder that translation may begin with an imperfect source transcript.

Prepare the source before translating

1. Approve the original transcript. Don't send an uncorrected draft into translation.
2. Create a terminology glossary. Preserve proper nouns, product names and specialist terms consistently.
3. Select priority languages. Use audience information and requests to decide where review effort belongs.
4. Assign fluent review. Check calls to action, numbers, prices, idioms and culturally sensitive wording in context.

A reviewer needs the video, not just isolated sentences. A phrase that sounds natural on its own may identify the wrong object or describe an action happening off-screen.

Give translation review a dedicated place in your editing workflow.

Test a complete video before scaling. Generate its captions, fix the consequential errors, check the exported result and approve any translated tracks. That gives you a repeatable process, not just a larger collection of unchecked files.

Taja AI turns long-form videos into captioned short clips and other repurposed content. Visit Taja AI to start free without a credit card, then use a video you already have to test the captioning and review process before preparing your next batch.