Captions
Ask “add captions” and Valmera writes word-timed captions from your transcript, styles them, keeps them clear of faces and burns them into the video. Change the look, highlight words, fix a spelling or hide captions during a section with one sentence. Captions are burned in; there’s no SRT or VTT file.
Add captions
- Use footage with speech
Captions are written from the word-level transcript, so upload a video with speech and let analysis finish.
- Ask for captions
Type “Add captions”, or name a style: “Add karaoke captions in yellow.” Without a style, Valmera picks one that suits the format, pace and brief.
- Review the preview
Check names and numbers, timing around cuts, and that captions stay clear of faces and on-screen text.
- Fix and restyle
Ask for spelling fixes (“spell it ‘Valmera’”) or style changes (“smaller, one line”). Text and timing stay as they are when you change the look.
- Export
Captions are burned into the exported MP4, in every frame where they appear in the preview.
- Add captions to the whole video.
- Add podcast-style captions and highlight the numbers.
- Add karaoke captions in yellow, two words at a time.
- Add clean white captions near the bottom, one line at a time.
Type these in Studio chat, or ask Claude, ChatGPT, Cursor or another assistant connected to Valmera’s MCP server. More in the prompt library.
Before it reports back, the agent audits the caption track (late first captions, missed words, overlaps, too many lines) and looks at rendered frames, so most timing and layout problems are fixed before you see them.
Caption styles
| Style | Looks like | Good for |
|---|---|---|
| clean | White, complete short phrases, size-led hierarchy | A safe default for most videos |
| documentary | Restrained subtitles on a translucent panel | Interviews, explainers |
| broadcast | Left-aligned, news-style lower third | News and explainer formats |
| reels | Tight two-line layout with a highlighted hero word | Short-form video |
| podcast | Bold words land as spoken; key words light up; numbers render big | Podcasts, talking heads |
| beast | All-caps impact font; the spoken word pops in the accent color | High-energy YouTube |
| karaoke | An accent box follows each spoken word | Shorts, music, energetic clips |
| spotlight | One glowing word at a time, centered | Hype edits, fast talkers |
| elegant | Calm lower third with serif-italic accents | Interviews, premium brands |
| lyric | Phrase-led typography | Music, quotes |
| stacked, iridescent, chrome, editorial, fashion, luxe, impact, retro, neon | Composed one-to-two-word looks with lines of different sizes | Stylized reels |
| classic | Plain subtitles | Simple, unstyled captions |
Change the style
Style changes keep the text and timing. Change one property or several in a sentence:
| Property | Options | Example request |
|---|---|---|
| Preset | Any style in the table above | “Switch to the elegant style” |
| Color and highlight | Any color for text and accent | “White text with a red highlight” |
| Size | S, M, L, XL, or a precise scale from 0.5× to 3× | “Make the captions 30% bigger” |
| Position | Bottom, middle, top, or an exact height | “Move the captions to the top” |
| Words per caption | 1 to 16, or one line at a time | “Two words at a time, single line” |
| Font and case | Bundled fonts; uppercase on or off | “Anton, all caps” |
| Animation | none, fade, pop, slide up, punch, blur in, whip, flash, rise, drop, elastic, bounce, swing, zoom blur | “Use a pop animation” or “no animation” |
| Finishing | Outline, shadow, backing panel, letter spacing, alignment | “Add a dark backing panel” |
Setting an explicit position locks captions to that band for the whole video. Leave position unset and Valmera moves captions shot by shot to avoid faces and on-screen text.
Highlight key words
Styles with emphasis highlight a few important words automatically, chosen from the words you stress, numbers and key terms in what you kept. Choose them yourself or turn emphasis off:
- Highlight “free”, “today” and “50%” in the captions.
- Make the highlighted words bigger instead of a different color.
- No highlighted words.
Fix a misspelled word
Ask in chat
“In the captions, change ‘Val Mira’ to ‘Valmera’.” Valmera replaces it everywhere in the captions, or in one place if you give the time. Fixes can change capitalization, punctuation and the number of words. The audio isn’t changed.
Correct the transcript
In the transcript panel, click the pencil on the sentence (Edit transcript line), correct it, then press Enter or Save (Esc cancels). A line holds up to 400 characters. The correction is saved for that video, so later captions and edits use it, and Studio re-renders the preview when the captions come from the transcript.
Captions you dictated yourself (see below) aren’t tied to the transcript, so change their text directly in chat.
Hide captions or write your own
Captions are skipped automatically over title cards and inserted clips, and designed text hides them while it’s on screen so words never print twice. To hide them yourself, give the times. To show text that isn’t in the speech, dictate it.
- Hide the captions from 0:42 to 0:50.
- Turn the captions back on everywhere.
- Add the caption “CHAPTER ONE” from 0:00 to 0:02.5 in the beast style.
- Remove all captions.
Type these in Studio chat, or ask Claude, ChatGPT, Cursor or another assistant connected to Valmera’s MCP server. More in the prompt library.
Placement and aspect ratios
Multi-word styles sit near the bottom, clear of the face; one-word-at-a-time styles such as spotlight belong mid-frame. When you reframe, the caption layout is redrawn for the new frame. Still watch the busiest shot in each ratio you export. For on-screen titles and lower thirds, see text and overlays.
Troubleshooting
| Symptom | Fix |
|---|---|
| A name or term is misspelled | Ask “change ‘Val Mira’ to ‘Valmera’ in the captions”, or fix the line with the pencil in the transcript panel. |
| Captions cover a face or a graphic | Ask to make them smaller or move them. If you locked a position earlier, ask to let captions move by shot again. |
| A caption appears late or early | Give the moment: “the caption at 0:42 shows up late.” The agent checks the word timing and fixes it. |
| Part of the video has no captions | That part is an inserted clip, a title card, or inside a caption-hidden window. Dictate captions there if you need them. |
| Two layers of captions | The footage has subtitles burned in. Ask to remove the old subtitles before adding new captions. |
| Too many words on screen | Ask for fewer words per caption or a single line. |
| A correction doesn’t show | Wait for the new preview. A file you downloaded earlier never changes; export again. |
Limits
- Captions are burned into the picture; there’s no SRT or VTT import or export, and viewers can’t turn them off.
- No custom font uploads; choose from the bundled fonts.
- Captions cover speech in the main video. Speech inside inserted clips isn’t transcribed.
- English is the best-tested transcription language.
Tools behind this page: add_captions set_caption_style set_caption_fixes set_caption_mutes audit_captions suggest_emphasis erase_burned_text · Full toolset reference
Frequently asked questions
How do I add captions in Valmera?
Upload a video with speech, wait for analysis, and ask “Add captions” or name a style such as “Add karaoke captions.” Valmera writes word-timed captions from the transcript, places them clear of faces, and burns them into the exported MP4.
Which caption styles are available?
Presets include clean, documentary, broadcast, reels, podcast, beast, karaoke, spotlight, elegant, lyric, stacked, iridescent, chrome, editorial, fashion, luxe, impact, retro, neon and classic. Any style can be adjusted for color, highlight color, size, position, words per caption, font and animation.
How accurate are Valmera's captions?
Captions come from a word-level speech transcript, so they match what was said and when. Names, brand terms and jargon are the usual errors; fix them with one request or in the transcript panel. English is the best-tested language. Before reporting back, the agent audits the caption track for late captions, missed words and overlaps.
Can I export SRT or VTT files?
No. Captions are burned into the video. There is no SRT or VTT import or export.
Can I use my own font?
Custom font uploads aren't supported. Choose from the bundled fonts: Inter Display (Black, ExtraBold, Bold), Anton, Bebas Neue, Archivo Black, Poppins Black, Syne ExtraBold, Playfair Display Black, Instrument Serif, DM Serif Display, Montserrat and Plus Jakarta Sans ExtraBold.
Do captions stay in sync after I cut the video?
Yes. Transcript captions follow the words you keep, so cuts, speed changes and inserts move them automatically. Check a few cut points in the preview.
Can I highlight the word being spoken, karaoke style?
Yes. Ask for karaoke captions (an accent box follows each spoken word) or a style like podcast, reels or beast where spoken and key words light up. You can set the highlight color and choose which words are emphasized.