How to add a freeze frame with text with Claude

Make input.mp4 open on a freeze frame: hold the first frame for 1.2 seconds with "You need to see this" in black on a white rounded box near the top, then play the clip normally with the text gone. Delay the audio by 1.2 seconds so it stays in sync. Use Python with Pillow to draw the text box and ffmpeg for the rest (don't rely on ffmpeg's drawtext filter), and save it as output.mp4.

#!/usr/bin/env python3
"""Freeze frame with text: the first frame holds still for 1.2 seconds under a headline, then the clip plays.

Usage: python3 freeze-frame-text.py input.mp4 output.mp4
The output is 1.2 seconds longer than the input; the sound waits for the freeze so it stays in sync.
Needs ffmpeg and Pillow (pip install pillow).
"""
import json
import subprocess
import sys
import tempfile
from pathlib import Path

from PIL import Image, ImageDraw, ImageFont

TEXT = "You need to see this"
FREEZE = 1.2     # seconds the first frame holds
SIZE = 0.085     # text height as a share of the frame width
Y = 0.16         # top of the text box (0 = top, 1 = bottom)
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 text_card(w):
    """The headline in black on a white rounded box, wrapped at 62% of the frame width so the lines come out even."""
    font = ImageFont.truetype(FONT_FILE, round(w * SIZE))
    lines, line = [], ""
    for word in TEXT.split():
        trial = f"{line} {word}".strip()
        if font.getlength(trial) <= w * 0.62 or not line:
            line = trial
        else:
            lines.append(line)
            line = word
    lines.append(line)
    pad, step = round(w * 0.04), round(font.size * 1.2)
    card = Image.new("RGBA", (round(max(font.getlength(l) for l in lines)) + 2 * pad, step * len(lines) + 2 * pad))
    draw = ImageDraw.Draw(card)
    draw.rounded_rectangle([0, 0, card.width - 1, card.height - 1], radius=pad, fill="white")
    for i, l in enumerate(lines):
        draw.text((card.width / 2, pad + step * i + step / 2), l, font=font, anchor="mm", fill="black")
    return card


src, out = sys.argv[1:3]
info = json.loads(subprocess.check_output(["ffprobe", "-v", "error", "-show_entries", "stream=codec_type,width", "-of", "json", src]))
width = next(s["width"] for s in info["streams"] if s["codec_type"] == "video")
has_audio = any(s["codec_type"] == "audio" for s in info["streams"])

with tempfile.TemporaryDirectory() as tmp:
    card_png = str(Path(tmp) / "card.png")
    text_card(width).save(card_png)
    graph = (f"[0:v]tpad=start_duration={FREEZE}:start_mode=clone[v];"
             f"[v][1:v]overlay=x=(W-w)/2:y=H*{Y}:enable='lt(t,{FREEZE})',format=yuv420p[out]")
    maps = ["-map", "[out]"]
    if has_audio:
        graph += f";[0:a]adelay={round(FREEZE * 1000)}:all=1[a]"
        maps += ["-map", "[a]"]
    subprocess.run(["ffmpeg", "-y", "-v", "error", "-i", src, "-i", card_png, "-filter_complex", graph, *maps,
                    "-c:v", "libx264", "-crf", "18", "-c:a", "aac", out], check=True)