hermoso
Login Sign up
Blog

Captions and end cards that look exactly how you describe them

They stay off unless you ask. When you do, say "big serif title with a small italic line" or "bold outlined caps" and that is the look that lands on your video.

By the Hermoso team·Launched September 11, 2026·Updated September 12, 2026·4 min read

Some ads want words on screen: a bold hook in the first second, a product name that sticks, a line at the end that says where to buy. Plenty of others look better with nothing at all, especially the phone-filmed videos that feel at home in a feed.

So Hermoso now does both properly. Captions and end cards stay off unless you ask for them. And when you ask, you get to say what they look like.

Off by default, on every kind of video

Every recipe renders clean unless you say otherwise: UGC, product demos, explainers and clones of other videos. No surprise subtitles and no logo card bolted onto the end. You decide, video by video.

Describe the look

In the Studio, just say it. Add captions in a big elegant serif with a small italic line under the title. Bold white caps with a black outline, near the bottom. A handwritten note, tilted a little. Hermoso turns that into a style and uses it for the captions and the end card together, so the two match.

Six starting looks cover most of what people ask for:

  • Editorial: a large elegant serif title in the middle of the frame with a small italic line underneath, no box.
  • Bold: condensed caps with a black outline that read on any footage.
  • Minimal: small, clean and out of the way.
  • Handwritten: a tilted marker note.
  • Boxed: text on a solid block of colour.
  • Pill: the rounded caption bubble.

Then tweak anything you like: the font (sans, serif, elegant, condensed or handwritten), the weight and size, the text colour, a box colour or no box at all, where it sits (top, center, lower or bottom), the letter case, italic, outline, shadow, a tilt of up to 12 degrees either way, and the end card's background colour.

Two lines from one caption

Write a caption as TITLE · small line and the part after the dot becomes the smaller second line. That is how you get the magazine look: a big product name with a quiet line of detail beneath it.

Every letter is set in real type

The words are typeset from real font files and laid over the finished video, not drawn by the video model. So your product name comes out spelled exactly the way you wrote it, in the font you picked, every time.

Ideas to try

  • The product reveal. Editorial style, the product name as the title and one quiet line of detail under it, held for a beat when the product lands in frame.
  • The testimonial quote. Handwritten, tilted slightly, placed lower third over a customer talking to camera, so it reads like a note someone scribbled.
  • The sound-off hook. Bold outlined caps at the top of the frame for the first two seconds, for the people scrolling with the volume down.
  • The launch date end card. Boxed, in your brand colour, with the date and where to buy.

Not sure which suits a video? Ask the Studio for two versions with different looks and pick the one you like, or run both and let the numbers decide.

Your wordmark on the close, when you want it

There is a separate switch for your brand wordmark and tagline over the last few seconds. Like captions and the end card, it only appears when you ask.

From your AI agent

Agents get the same controls on render_ad:

  • captions: true and endCard: true turn them on.
  • textStyle takes a preset name like "editorial", or an object with preset plus any overrides such as font, color, position or tilt.
  • lockup: true adds the wordmark and tagline.

A mistyped style field is caught before anything renders, with the field named, so an agent can fix it and go again without spending a render.

It pairs nicely with video cloning: clone a video that had a big title card, ask for captions in the same spirit, and describe the type.

Questions people ask

Are captions on by default?

No. Captions, end cards and the brand wordmark are all off unless you ask for them, on every kind of video.

What caption styles are there?

Six presets: editorial, bold, minimal, handwritten, boxed and pill. On top of any of them you can change the font, size, colour, box, position, case, italic, outline, shadow and tilt.

Can I use my brand colours?

Yes. Set the text colour, the box colour and the end card background with hex codes.

Will the words be spelled correctly?

Yes. Captions are typeset from real font files and composited onto the finished video, so they come out exactly as written.

Can an AI agent set the caption style?

Yes. Pass textStyle on render_ad with a preset name or an object of overrides, together with captions: true or endCard: true.