Free AI Subtitle Generators: Where to Caption Videos at No Cost (2026)

The short version: CapCut, YouTube Studio, Canva, and the free tier of Kapwing will all caption your video for nothing right now, and for most short social clips they're good enough. The catch is that "free" means you're doing the editing, not the AI, because every single one of these tools gets names, brands, and accents wrong often enough that you have to check every line before you hit publish.

Why I bothered testing every one of these instead of just picking one

Last year I was cutting a launch video for a client, a two-minute clip announcing a new service, meant for LinkedIn. My editor was on holiday. The client wanted it live in three hours. I didn't have £40 to £80 spare to send it to a freelancer on a caption gig, and I certainly didn't have time to wait for one to come back to me.

So I ran it through four different free tools that afternoon, back to back, on the same file, just to see what happened. That's the test most "best free subtitle tool" listicles skip. They tell you a tool exists. They don't tell you what it does to your actual footage, with your actual client's actual company name in it.

Here's what happened: three out of four tools spelled the client's brand name wrong. One turned "Bullock" into "Bullick." One dropped a whole sentence during a moment where the client leaned forward and spoke faster. None of that is a knock on the technology being bad, it's a reminder that auto-captioning is a first draft generator, not a finished product, no matter how good the demo video on the tool's homepage looks.

The free tools that are worth your time in 2026

These are the ones I've used on real client work, not ones I've read about.

• CapCut. Free on desktop and mobile, and this is my go-to for anything short-form. Auto captions are quick, the accuracy on a clear British or American accent is strong, and you can style the text (bold, highlighted word, karaoke-style) without paying anything. The catch: heavier accents and background music both drop accuracy noticeably, and 4K exports sometimes push you toward the paid tier.
• YouTube Studio. Every video you upload to YouTube gets free auto-captions, no separate tool needed. It's unlimited and it's built in. The problem is the accuracy sits somewhere around 70 to 85 percent depending on audio quality and accent, and YouTube will not tell you which words it's guessed at, so you have to review the whole transcript manually before you publish, which most people never do.
• Canva. If you're already using Canva for social graphics, its video editor has a free auto-subtitle feature built into the timeline. It's not as polished as CapCut's styling options but it's convenient if you're already in there designing a thumbnail.
• Kapwing. The free plan caps you at short exports with a watermark, which is fine for testing but not for anything client-facing. Useful if you want to try the auto-transcribe feature before deciding whether to pay.
• VEED. Free plan is limited by minutes per month and comes with a watermark on export, but the auto-caption accuracy is decent and the translation feature (turning your English captions into other languages) is one of the better free implementations I've tried.
• Descript. The free tier gives you a set amount of transcription time each month. It's built for podcasters and long-form video editors rather than quick social clips, but if you're captioning a webinar or a 20-minute interview, this is the one I'd reach for first because it lets you edit the video by editing the text, which is a different way of working entirely and saves a huge amount of time once you're used to it.

If you're building out a wider set of tools rather than hunting for one-off wins, it's worth looking at how these fit into the AI tool stack most business owners need rather than collecting apps one at a time as problems come up.

Step by step: captioning a video for free in under ten minutes

This is the process I used on that launch video, and it's the same one I'd use today.

1. Upload the raw file. Most tools accept MP4 or MOV up to a few hundred MB on the free tier.
2. Run the auto-transcribe. This takes anywhere from 30 seconds to 3 minutes depending on video length.
3. Read every single line against the audio. Not skim it, read it, word by word, while the video plays. This is the step nearly everyone skips and it's the one that determines whether your captions embarrass you.
4. Fix proper nouns first. Names, brands, product names, technical terms. This is where auto-captions fail most often and most visibly.
5. Check the numbers. AI transcription tools frequently mishear spoken numbers, turning "fifteen percent" into "fifty percent" or dropping decimal points entirely. If your video mentions pricing, stats, or dates, check every single one manually.
6. Style the text. Font size big enough to read on a phone screen without zooming, high contrast against the background, positioned so it doesn't sit over a logo or a face.
7. Export and watch it on mute. If you can follow the story with the sound off, you're done. If you can't, go back to step three.

That's it. Ten minutes for a two-minute video, once you know what you're checking for. The first time it'll take you closer to twenty because you'll be second-guessing yourself on every line.

Why captions matter more than most people think

This isn't a nice-to-have anymore. A frequently cited industry stat, going back to a Verto Analytics study reported by Forbes and others, found that around 85 percent of Facebook video is watched with the sound off. LinkedIn autoplay works the same way. If your video has no captions, the vast majority of people scrolling past it will never hear a word you said, and most will keep scrolling before they turn the sound on to find out if it's worth their time.

On top of that, captions aren't just for silent scrolling. They matter for viewers who are deaf or hard of hearing, for anyone watching in a language that isn't their first, and for anyone watching in a noisy office or a quiet train carriage. Skipping captions isn't a small oversight, it's cutting off a huge chunk of your potential audience before the video even starts.

The bit that most guides on this topic leave out

Here's the uncomfortable truth: "free" doesn't mean free, it means the cost has moved from your wallet to your time, and most people don't budget for that. I've watched business owners spend forty minutes fighting with a free tool's clunky editor to fix three captions, when a fifteen-minute call with someone who does this daily would have sorted the whole video, styling and all, in the time it took them to open the app.

Free is the right call when the video is low-stakes and the mistakes are forgivable, a quick behind-the-scenes clip, an internal team update, a casual Instagram story. Free stops being the right call the moment the video represents your brand to a new audience, a client pitch, a paid ad, a launch video that's going in front of thousands of people who've never heard of you before. That's when one mangled brand name or one wrong number in the captions does more damage than the money you saved by not paying for a proper tool or a proper editor.

There's also a quieter issue nobody talks about: most free auto-caption tools use your audio and video to train their own models. Read the terms before you upload anything with client information, unreleased product names, or anything under an NDA. It's rarely spelled out clearly, and it should be one of the first things you check, not the last.

When it's worth paying, even a little

If you're producing video regularly rather than the occasional clip, the maths flips fast. A single Kapwing or VEED subscription runs somewhere in the £10 to £25 a month range and removes the watermark, the export caps, and usually improves accuracy on background noise and accents. If you're doing more than four or five videos a month, that's cheaper than the hours you'll burn wrestling with free tier limits, and it looks a lot more professional to a client than a watermark sitting in the corner of every frame.

If you're producing video in more than one language, this is where free tools fall apart fastest. Auto-translation on the free tiers is workable for casual content but rough for anything client-facing, and this is exactly where a proper look at AI auto captions and translation as part of your wider content strategy pays for itself, because getting captions right in more than one market is a completely different job to captioning a single English-language clip.

And if you're already editing full videos rather than just adding captions to something a phone shot, it's worth pairing your subtitle tool with a proper editor. There's a good rundown of free AI video editor tools worth knowing about that cover the cutting, resizing, and polishing side, which is usually the bigger time sink anyway.

What I'd tell a small business owner starting today

Start with CapCut if your content is short-form and social. Start with YouTube Studio's built-in captions if you're already uploading to YouTube and don't want another app in the mix. Start with Descript if you're doing longer interview or podcast-style video and want to edit by text rather than timeline.

Whichever one you pick, budget the review time as part of the job, not as an optional extra. Ten minutes of checking every line against the audio is the difference between captions that make you look sharp and captions that make you look like you didn't bother to check your own video before putting your name on it.

Frequently asked questions

Which free AI subtitle generator is the most accurate?

CapCut and YouTube Studio's built-in captions tend to perform best on clear, unaccented English audio with minimal background noise. Accuracy drops on both once you add strong regional accents, overlapping speakers, or music under the dialogue, so treat any free tool's output as a first draft that still needs a manual check.

Do free caption tools work for languages other than English?

Most do, but accuracy and translation quality vary a lot between languages. VEED and Kapwing both offer basic translation on their free tiers, but for anything client-facing in a second language, it's worth reviewing every line with a native speaker before publishing rather than trusting the auto-translation outright.

Will free subtitle tools put a watermark on my video?

Some will, some won't. Kapwing and VEED both add watermarks and export limits on their free plans. CapCut and Canva generally don't, which is part of why they're the better starting point for anything you're planning to share publicly under your own brand.

Is it safe to upload client videos to free AI caption tools?

Check the tool's terms before uploading anything confidential. Many free tools use uploaded audio and video to train their own models, and this is rarely stated clearly on the main sign-up page. If a video contains client names, unreleased products, or anything covered by an NDA, read the data policy first or use a paid tier with stricter data handling.