--- title: "How to add dual-language captions with Claude (tested prompt and script)" description: "Each spoken phrase appears in white near the bottom with its Spanish translation in smaller yellow type right under it, both on one dark strip. Install one skill and ask Claude Code, or run the Python script we tested yourself." url: https://www.recharm.com/effects/dual-language-captions --- # How to add dual-language captions with Claude Each spoken phrase appears in white near the bottom with its Spanish translation in smaller yellow type right under it, both on one dark strip. Made by the script on this page, from our own test clip. ## Get this effect 1. ### Switch to Code in your [Claude Desktop App](https://claude.com/download) 2. ### Install the skill (once) ``` Install the Recharm video effects skill: download https://recharm-growth-staging.s3.us-east-2.amazonaws.com/website/downloads/recharm-video-effects.zip and unzip it into ~/.claude/skills ``` 3. ### Ask for the effect ``` Add dual-language captions to input.mp4 Translation: ["Este sérum limpió mi", "piel en dos semanas"] ``` ## Several videos, several effects, one ask Put your clips in the folder, mention them with @, and ask for everything at once. Claude runs one tested script after another on each clip, then joins them. Add dual-language captions to **@clip1.mp4**, put a sale badge on **@clip2.mp4**, give **@clip3.mp4** a VHS look, then join all three with crossfades. ## Settings you can change Ask Claude to change any of these, or edit them at the top of the script. | Setting | Default | What it does | | ------------- | ------------------------------------------------- | -------------------------------------------------------------------------- | | `TRANSLATION` | `["Este sérum limpió mi", "piel en dos semanas"]` | One line in the second language for each spoken phrase, in order. | | `CHUNK` | `4` | Words per phrase. | | `SIZE` | `0.064` | Spoken line's text height as a share of the frame width. | | `SECOND_SIZE` | `0.054` | Translation's text height as a share of the frame width. | | `Y` | `0.80` | Where the spoken line's center sits, from top (0) to bottom (1). | | `GAP` | `0.055` | Distance down to the translation's center, as a share of the frame height. | | `SECOND_TEXT` | `(255, 214, 10)` | Translation color. | | `STRIP` | `(0, 0, 0, 170)` | Color and opacity of the strip behind both lines. | ## Without the skill Don’t want to install the skill? Paste this prompt into Claude Code and Claude writes the script itself. ``` Add dual-language captions to input.mp4. Transcribe the speech with faster-whisper (model base.en, word timestamps), group the words four at a time, and translate each group into Spanish. Show each group from its first word's start to its last word's end and nothing after the last word: the English in white Arial Bold 6.4% of the frame width tall, centered horizontally with its center 80% of the way down the frame, and the Spanish in yellow (255, 214, 10) Arial Bold 5.4% of the frame width tall, centered 5.5% of the frame height below it. Put both on one black strip at 67% opacity (alpha 170) with rounded corners, reaching 3% of the frame width past the wider line on each side and 2.1% above the English and below the Spanish. Use Python with Pillow to draw the text (don't use ffmpeg's drawtext or subtitles filter) and ffmpeg to put it on the video, keep the audio, and save it as output.mp4. ``` ## Run it without Claude This is the exact script we tested. Save it as `dual-language-captions.py` and run: ``` python3 dual-language-captions.py input.mp4 output.mp4 ``` ``` #!/usr/bin/env python3 """Dual-language captions: the spoken phrase in white with its translation in yellow underneath, on a dark strip. Usage: python3 dual-language-captions.py input.mp4 output.mp4 Whisper transcribes the speech and splits it into phrases of CHUNK words. TRANSLATION holds one line per phrase in the second language (ask Claude to translate them); if the count doesn't match, the script prints the phrases and stops so you can fill it in. Needs ffmpeg, Pillow and faster-whisper (pip install pillow faster-whisper). """ import json import math import subprocess import sys from pathlib import Path from faster_whisper import WhisperModel from PIL import Image, ImageDraw, ImageFont TRANSLATION = ["Este sérum limpió mi", # one line per spoken phrase, in order "piel en dos semanas"] CHUNK = 4 # words per phrase SIZE = 0.064 # spoken line's text height as a share of the frame width SECOND_SIZE = 0.054 # translation's text height as a share of the frame width Y = 0.80 # where the spoken line's center sits (0 = top, 1 = bottom) GAP = 0.055 # distance down to the translation's center, as a share of the frame height TEXT = "white" # spoken line color SECOND_TEXT = (255, 214, 10) # translation color STRIP = (0, 0, 0, 170) # color and opacity (0-255) of the strip behind both lines PAD = 0.03 # space around the text inside the strip, as a share of the frame width FONTS = ["/System/Library/Fonts/Supplemental/Arial Bold.ttf", # macOS "/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf", # Linux "C:/Windows/Fonts/arialbd.ttf"] # Windows FONT_FILE = next((f for f in FONTS if Path(f).exists()), None) or sys.exit("No bold font found: add one to FONTS") def phrases(src): """[(start, end, text)] for every CHUNK spoken words, from Whisper's word timings.""" model = WhisperModel("base.en", device="cpu", compute_type="int8") segments, _ = model.transcribe(src, word_timestamps=True) words = [word for segment in segments for word in segment.words] groups = [words[i:i + CHUNK] for i in range(0, len(words), CHUNK)] return [(g[0].start, g[-1].end, " ".join(word.word.strip() for word in g)) for g in groups] def caption(spoken, translated): """Both lines drawn once, centered, on one rounded strip.""" img = Image.new("RGBA", (w, h)) draw = ImageDraw.Draw(img) top, bottom = (w / 2, h * Y), (w / 2, h * (Y + GAP)) a = draw.textbbox(top, spoken, font=font, anchor="mm") b = draw.textbbox(bottom, translated, font=second_font, anchor="mm") pad = w * PAD draw.rounded_rectangle([min(a[0], b[0]) - pad, a[1] - pad * 0.7, max(a[2], b[2]) + pad, b[3] + pad * 0.7], radius=pad * 0.6, fill=STRIP) draw.text(top, spoken, font=font, anchor="mm", fill=TEXT) draw.text(bottom, translated, font=second_font, anchor="mm", fill=SECOND_TEXT) return img def overlay(t): for i, (start, end, _) in enumerate(cues): if start <= t < end: return drawn[i] return blank def probe(path): info = json.loads(subprocess.check_output(["ffprobe", "-v", "error", "-select_streams", "v:0", "-show_entries", "stream=width,height,r_frame_rate:format=duration", "-of", "json", path])) s = info["streams"][0] num, den = s["r_frame_rate"].split("/") return s["width"], s["height"], s["r_frame_rate"], float(num) / float(den), float(info["format"]["duration"]) def render(src, out): """Draw overlay(t) on every frame of src and save to out, keeping the audio.""" ff = subprocess.Popen(["ffmpeg", "-y", "-v", "error", "-i", src, "-f", "rawvideo", "-pix_fmt", "rgba", "-s", f"{w}x{h}", "-r", rate, "-i", "-", "-filter_complex", "[0:v][1:v]overlay=shortest=1,format=yuv420p[v]", "-map", "[v]", "-map", "0:a?", "-c:v", "libx264", "-crf", "18", "-c:a", "copy", out], stdin=subprocess.PIPE) for i in range(math.ceil(duration * fps)): ff.stdin.write(overlay(i / fps).tobytes()) ff.stdin.close() if ff.wait(): sys.exit("ffmpeg failed") src, out = sys.argv[1:3] cues = phrases(src) if len(cues) != len(TRANSLATION): sys.exit("Set TRANSLATION to one line for each of these phrases:\n" + "\n".join(text for _, _, text in cues)) w, h, rate, fps, duration = probe(src) font = ImageFont.truetype(FONT_FILE, round(w * SIZE)) second_font = ImageFont.truetype(FONT_FILE, round(w * SECOND_SIZE)) blank = Image.new("RGBA", (w, h)) drawn = [caption(text, translated) for (_, _, text), translated in zip(cues, TRANSLATION)] render(src, out) ``` ## More captions and text - [AI chat reply animation **AI chat reply animation**](https://www.recharm.com/effects/ai-chat-typing) - [Animated bar chart **Animated bar chart**](https://www.recharm.com/effects/animated-bar-chart) - [Animated checklist **Animated checklist**](https://www.recharm.com/effects/animated-checklist) - [Animated line graph **Animated line graph**](https://www.recharm.com/effects/animated-line-graph) - [Animated pie chart **Animated pie chart**](https://www.recharm.com/effects/animated-pie-chart) ## Find 3x _winning ads_ Recharm cuts your ad footage into tagged, searchable clips: hooks, testimonials, product shots. Find the shots that win, add the effect with Claude, and ship 3x more ad edits. [Start 14-day free trial](https://www.recharm.com/get-started)