How to add highlighted word captions with Claude

Add captions to input.mp4 using the word timings in words.json. Show the words three at a time in capitals, centered 70% of the way down, white bold text with a black outline, and turn the word being spoken yellow (255, 212, 0). Each group stays up from its first word's start to its last word's end. Use Python with Pillow to draw it and ffmpeg to put it on the video (don't rely on ffmpeg's drawtext filter), keep the audio, and save it as output.mp4.

#!/usr/bin/env python3
"""Highlighted word captions: short phrases on screen, with the word being spoken lit up in yellow.

Usage: python3 highlighted-word-captions.py input.mp4 words.json output.mp4
words.json holds word timings, for example from Whisper:
  [{"word": "This", "start": 0.30, "end": 0.69}, {"word": "serum", "start": 0.77, "end": 1.13}, ...]
Needs ffmpeg and Pillow (pip install pillow).
"""
import json
import math
import subprocess
import sys
from pathlib import Path

from PIL import Image, ImageDraw, ImageFont

SIZE = 0.085                # text height as a share of the frame width (shrinks to fit long phrases)
Y = 0.70                    # where the line sits (0 = top, 1 = bottom)
CHUNK = 3                   # words on screen at once
HIGHLIGHT = (255, 212, 0)   # color of the word being spoken
STROKE = 0.01               # outline thickness 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")


fonts = {}


def fitted_font(texts):
    """The largest font up to SIZE that fits the phrase in 90% of the frame width."""
    size = round(w * SIZE)
    while size > 12:
        font = fonts.get(size) or fonts.setdefault(size, ImageFont.truetype(FONT_FILE, size))
        if sum(font.getlength(s) for s in texts) + font.getlength(" ") * (len(texts) - 1) <= w * 0.9:
            return font
        size -= 2
    return font


def overlay(t):
    img = Image.new("RGBA", (w, h))
    draw = ImageDraw.Draw(img)
    for group in (words[i:i + CHUNK] for i in range(0, len(words), CHUNK)):
        if not group[0]["start"] <= t < group[-1]["end"]:
            continue
        spoken = [word for word in group if word["start"] <= t][-1]  # the word being said, or the last one said
        texts = [word["word"].strip().upper() for word in group]
        font = fitted_font(texts)
        space = font.getlength(" ")
        x = (w - sum(font.getlength(s) for s in texts) - space * (len(texts) - 1)) / 2
        for word, s in zip(group, texts):
            draw.text((x, h * Y), s, font=font, anchor="lm", fill=HIGHLIGHT if word is spoken else "white",
                      stroke_width=round(w * STROKE), stroke_fill="black")
            x += font.getlength(s) + space
    return img


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, words_file, out = sys.argv[1:4]
words = json.loads(Path(words_file).read_text())
w, h, rate, fps, duration = probe(src)
render(src, out)