How to add captions to a video automatically
Add captions to a video automatically inside Claude — ask Polaris to caption your clip and it burns TikTok-native captions in, done in about a minute.
To add captions to a video automatically, connect Polaris to Claude and ask it in plain chat to caption your clip — Polaris transcribes the audio word-for-word, applies a style like TikTok, Hormozi, Beast, or Neon, and burns the text into the frame. A live panel fills in the finished, captioned video by itself in about a minute, ready to post. (You can do the same in the Polaris web app.)
What "automatic captions" actually means
Automatic captions use speech-to-text to listen to your video, write out every word, time each word to the audio, and render that text directly onto the video. There is no manual typing and no dragging timestamps around a timeline. The two things that matter for social video ads are accuracy (the words match what was said) and burn-in (the captions are part of the video file, not a separate subtitle track).
Burn-in is the important part for ecommerce. A separate .srt subtitle file gets stripped or ignored the moment your video hits TikTok, Reels, or Shopify. Burned-in captions travel with the video everywhere it goes, and they show up even when the viewer is scrolling with the sound off, which is how the large majority of feed video gets watched.
How to add captions to a video automatically with Polaris
Here is the full workflow, start to finish:
- Connect Polaris to Claude once. Add the MCP connector (polarisugc.com/api/mcp) in Claude's connectors and authorize it with OAuth. From then on, Polaris lives right inside your chat — there is no dashboard to babysit.
- Ask Claude to caption your clip. In plain chat, attach the video you want captioned — a product demo, a talking-head UGC ad, or a video you just generated inside Polaris — and tell it to burn captions in.
- Pick a caption style by tapping a card. Choose a TikTok-native preset — TikTok, Hormozi, Beast, or Neon — from the option cards in the panel. The style sets the font, color, highlight, and animation feel.
- Watch it render and download. A live Polaris panel transcribes the audio word-accurate, times each word to the moment it is spoken, burns the captions in, and hands you a finished file (about 20-180 seconds) ready to post. You do not type anything.
Because the captions are burned in, the exported video is final. There is no extra "attach subtitles" step on the platform side, and nothing to re-sync after you post.
Pick a caption style that fits the ad
Different styles pull attention in different ways. All four are built to feel native to short-form feeds, and all four are burned in and word-timed. Here is how we think about each:
| Style | Look | Best for |
|---|---|---|
| TikTok | Clean, familiar, platform-native captions | Everyday UGC and organic-feeling ads |
| Hormozi | Bold, punchy, word-by-word highlight pops | High-energy hooks and offer-driven ads |
| Beast | Big, loud, high-contrast emphasis | Fast-cut, retention-focused videos |
| Neon | Glowing, colorful accents | Standing out on a busy, visual feed |
If you are not sure, we usually start with TikTok for a natural look or Hormozi when the goal is to drive a specific offer. You can caption the same video in more than one style and test which one holds attention longer.
Tips for clean, accurate captions
- Feed it clear audio. Word accuracy tracks with audio quality. A clean voiceover or a well-recorded talking-head transcribes more cleanly than a noisy room.
- Match the style to the message. A calm, premium product benefits from the clean TikTok look; a punchy discount hook lands harder with Hormozi or Beast.
- Keep the hook in the first seconds. Captions earn their keep in the opening moments when a viewer decides to stay or scroll, so make sure your strongest line is spoken early.
- Design for sound-off. Assume most viewers never turn the sound on. If the captions alone tell the story, the ad works in the feed.
Add captions inside Claude with the MCP connector
Because Polaris runs inside Claude, captioning is part of the same conversation where you make the ad. Ask Claude to generate the video ad inside Claude, then caption it in the same chat — Polaris transcribes, applies the style you ask for, and returns a captioned file without leaving the conversation. That takes you from raw clip to captioned, ready-to-post ad in one place. Prefer a prompt-first composer? The Polaris web app runs the same engine.
Why teams pick Polaris
Polaris is an AI ad studio built only for ecommerce. Auto-Captions is one piece of a workflow designed to take you from a product image or prompt all the way to a finished ad:
- Captions built for ads, not documents. Four-plus TikTok-native styles, transcribed word-accurate and burned straight into the video.
- The whole ad in one place. Generate UGC-style video ads from an image or prompt using multiple top models chosen per shot (Google Veo, Kling, Seedance, Nano Banana), create product-shot images, then caption the result.
- Recreate winning ads. Paste a reference ad and get your own version, then caption it in your chosen style.
- Works inside Claude. The MCP connector lets you make and caption ads directly in chat.
- Credit-based pricing. Pro is $49/mo with 2,000 credits, and it gets cheaper on longer terms — $42/mo on a 3-month plan and $37/mo on 6 months.
The point is speed without the drop in quality. Captions that used to mean opening an editor, typing every line, and hand-timing the text now take a single request and a style pick.
Get started
Adding captions to a video automatically comes down to four steps: ask Claude, transcribe, pick a style, burn in. Polaris does all four for you inside the chat and returns a video that is ready for TikTok, Reels, and your Shopify store. Create your first ad or see pricing to get going.